Datasets:
ccdv
/
arxiv-summarization

Dataset card Data Studio Files Community
7
article
stringlengths
0
745k
abstract
stringlengths
0
92.7k
the quantum heisenberg xy model , which was first intensively studied by lieb et al in the early 1960s @xcite , has aroused new interests in recent years . it was taken to be a potential candidate model in the practical applications in quantum computation @xcite and quantum information @xcite . the heisenberg model is not only a theoretical model with many interesting physics in it , but also realized in recent years in laboratory in physical systems such as quantum dots @xcite , nuclear spins @xcite and electronic spins @xcite . many theoretical researchs @xcite had been devoted into the model to explore the important quantum properties , especially the correlation function @xcite , the entanglement @xcite and quantum phase transition @xcite in one dimensional systems . the studies show that the amount of the spatial correlation and entanglement between two spins in a spin chain ( with open end ) or a spin ring ( with periodical boundary condition ) can be modified by changing the magnitude of the external magnetic field and/or the interaction strength . and these two quantities shows singularity and obeys the scaling law in the vicinity of the quantum phase transition point of the system . + the quantum phase transition , which can take place at @xmath0 , is an interesting subject and intensively studied in the last decades @xcite . in this paper , we studied the quantum phase transitions of the xy model through the calculation of the correlation and the concurrence @xcite between two nearest - neighbor spins . by changing the magnetic field or coupling coefficient , we find that both of the two quantities have finite jumps at some critical points . the states of the system are carefully studied , the jumps are related to the level crossing @xcite of the ground state to the excited state . the change of the symmetry of the ground state @xcite is the course of the quantum phase transition . the rest of this paper is organized as follows . in section [ sec : hamiltonian ] , we introduce the hamiltonian for the heisenberg xy model in an external field and describe the computation procedures for the correlation and the concurrence of the subsystem ; detailed results and discussions are given in section [ sec : discussion ] ; the conclusion of our study is given in section [ sec : conclusion ] . the model studied here is a one dimensional spin-1/2 chain with periodical boundary condition ( pbc ) @xcite under a transverse isotropic magnetic field . the hamiltonian is : @xmath1 where @xmath2 is the number of spins , @xmath3 and @xmath4 from the pbc . @xmath5 and @xmath6 are the coupling strength between nearest - neighbor spins along the direction of @xmath7 and @xmath8 , respectively , @xmath9 is the amplitude of the magnetic field along @xmath10 direction . @xmath11 , @xmath12 and @xmath13 are the well - known pauli matrix : @xmath14 the parameters @xmath5 and @xmath6 may be represented by their average value @xmath15 and anisotropic parameter @xmath16 as : @xmath17 and the hamiltonian may be written as : @xmath18 + \sum_{i=1}^{n}b_z\sigma_z^{(i)},\ ] ] where @xmath19 ) : @xmath20 the spatial correlation @xcite between two spins , 1 and 2 , is defined by @xmath21 where @xmath22 can be represented explicitly in a matrix form in the product eigenspace of @xmath23 : @xmath24 and @xmath25 is the reduced density matrix of the spin 1 and 2 derived from the density matrix @xmath26 of the whole system . @xmath27 means to trace out all the spin degrees of freedom except spins 1 and 2 . in this study , we are interested in the ground state of the whole system , which is the lowest energy eigenstate @xmath28 of @xmath29 , so @xmath30 . + the concurrence of spins 1 and 2 is defined as @xcite : @xmath31 where @xmath32 are the square roots of the eigenvalues of the product matrix @xmath33 in decreasing order . where @xmath25 is the reduced density matrix of the two spins subsystem , and @xmath34 is constructed as @xmath35 . equation ( [ concurrence ] ) applies to both of the mixed state and pure state . the reduced density matrix may be write down implicitly as : @xmath36 then we get @xmath34 as the matrix below : @xmath37 if the bipartite quantum state @xmath25 is pure , it can always be written as : @xmath38 ) could be simplified to @xmath39 .the discontinuities in fig . [ fig : ssaf4 - 10 ] [ cols="^,^,^,^,^,^",options="header " , ] [ tal : disbz ] the effect of the magnitude of magnetic field on the phase transition was shown in figure [ fig : cor6jr0 ] , in which we plot the case of a @xmath40 spins system with three different magnetic fields , @xmath41 , @xmath42 and @xmath43 . the corresponding transition points of @xmath15 are listed in table [ tal : disbz ] . by simple calculation and comparison , we found that the transition values of @xmath15 for each transition is approximately proportional to the magnitude of the external field . + figure [ fig : cor6b ] plots the variation of correlation @xmath44 with the applied field @xmath9 for fixed coupling coefficient @xmath45 and different values of anisotropic parameter @xmath16 . the system sizes are @xmath46 . when @xmath47 , there are @xmath48 jumps for a 6 spins system as in the case of isotropic case . however , with the increase of anisotropy @xmath16 , the magnitude of the jumps becomes smaller and the transition points moves slightly to the direction @xmath49 , till finally disappears in the completely anisotropic case @xmath50 . since the @xmath50 corresponding to the ising case , where the internal number of freedom is 1 , the phase transition is different to the xy case where the internal number of freedom is 2 . our calculation shows that the cross over from @xmath51 to @xmath52 is continuous . it is also verified that there is no quantum phase transition in the quantum ising case . + now we turn to the concurrence between two spins , @xmath53 . the following results were obtained for a system of @xmath46 . figures [ fig : con6jr0 ] plot the variation of concurrence @xmath53 with the coupling strength @xmath15 in the isotropic case ( @xmath54 ) , the applied magnetic field fixed to three values , @xmath55 , @xmath56 and @xmath57 . from the figures we see the same phase transition behavior as that of the correlation functions . jumps are found at exactly the same places , which can also be checked in table [ tal : disbz ] , as for the correlation functions . figures [ fig : con6b ] gave the variation of @xmath53 with the magnitude of external field @xmath9 for @xmath58 different anisotropy @xmath16 . when @xmath59 , the jumps in concurrence are clearly observed while for larger @xmath60 s the magnitude of the jumps becomes smaller and smaller and turn to the smooth curve in the case of @xmath61 . + it is interesting to note that the physics behind the above phase transitions turns out to be very simple and clear , it is the result of level crossing of the ground state . to clarify this point , we turn to a detailed analysis of the ground states of the system under different sets of parameters . to be specific , we use @xmath62 and @xmath63 to represent the `` spin down '' and `` spin up '' state for each spin in the eigenspace of @xmath23 respectively . the state of the whole system of @xmath2 spins @xmath64 can be written as a linear combination of @xmath65 states and @xmath65 normalized coefficients : @xmath66 where @xmath67 stands for the direct product state @xmath68 . we diagonalized the hamiltonian and calculated the eigen function and the corresponding eigen energy . we take the coupling strength @xmath69 and anisotropy @xmath70 for simplicity . figure [ fig : esbz ] shows the variation of several states energy ( someone of them will be the ground state in some interval of @xmath71 ) with the magnitude of external field @xmath9 . it is clear from the figure that level crossing occurs at the positions where abrupt changes of the correlation function and the concurrence were observed . concretely , when @xmath72 , the ground state ( which has been normalized ) of the whole system is @xmath73 which is basically the combination of the states with antiferromagnetic order , that is , the states with half of the spins `` up '' and half of the spins `` down '' . when @xmath74 , the ground state changed to @xmath75 this is the state that 2 spins `` up '' and 4 spins `` down '' . in the course of increasing the applied field @xmath9 , the energy level corresponding to the state ( [ first ] ) increases and the energy level corresponding to the state ( [ second ] ) decreases . at the critical field , @xmath76 , the two levels meet and then crossed , then the ground state of the spin chain is a combination of the two level states . beyond the critical field the ground state becomes the state ( [ second ] ) . the cross of the ground state changed the symmetry of the ground state abruptly so that a quantum phase transition occurs . at about @xmath77 , another level cross take place and over that point , the ground state of the system becomes , @xmath78 this is a combination of states with 1 spins `` up '' and 5 spins `` down '' . and finally , when @xmath79 , the ground state is simply as @xmath80 which is the state with ferromagnetic order . we see from the above analysis that the quantum phase transition is a result of the level crossing of the ground state , and the level crossing is simply the result of the competition between the coupling strength @xmath15 , which favors antiferromagnetic order , and the applied field , which favors the ferromagnetic order . by increasing the applied field while keeping the coupling strength constant , the ground state changed from antiferromagnetic order to the ferromagnetic order in several steps , turns a spin from `` up '' to `` down '' at each step . so we observed that @xmath81 . this also explains the number of jumps in the correlation function and concurrence , which must be half of the whole number of spins . in this paper , we studied the spatial correlation @xmath82 and the degree of entanglement @xmath53 between two nearest - neighbor spins along a one - dimension spin ring in heisenberg xy model . we found that in the isotropic case with fixed external field @xmath9 , quantum phase transitions of both @xmath82 and @xmath53 were observed with the increment of coupling strength @xmath15 , the phase transitions will smoothed out as the anisotropy @xmath16 increased and disappears at the extremum value of @xmath50 . the transition is a direct result of the level crossing of the ground state of the whole system , in which the symmetry of the ground state changed abruptly at the crossing point . 100 e. lieb , t. schultz , and d. mattis , ann . phys.n.y . * 16 * , 407 ( 1961 ) s.b . zheng and g.c . guo , phys . lett . * 85 * , 2392 ( 2000 ) l. you and m.s . chapman , phys . a * 62 * , 052302 ( 2000 ) r. raussendorf and h.j . briegel , phys . rev . lett . * 86 * , 5188 ( 2001 ) a. imamoglu , d.d . awschalom , et . al . , phys . lett . * 83 * , 4204 ( 1999 ) g. burkard , d. loss and d.p . divincenzo phys . rev . b * 59 * , 2070 ( 1999 ) d. loss and d.p . divincenzo , phys . a * 57 * , 120 ( 1998 ) b.e . kane , nature ( london ) * 393 * 133 ( 1998 ) r. vrijen et al . , e - print quant - ph/9905096 k.m . oconnor , w.k . wootters , phys . a * 63 * 052302 ( 2001 ) x. wang , phys . a * 64 * 012313 ( 2001 ) s. liang , phys . lett * 64 * 1597 ( 1990 ) m.a . nielson and i.l . chuang , _ quantum computation and quantum information _ cambridge university press , cambridge , england ( 2000 ) s. sachdev , _ quantum phase transitions _ ( cambridge university press ) cambridge , england ( 2000 ) x. peng , j.f . du , and d. suter , phys . rev . a , * 71 * , 012307 ( 2005 ) h.t . quan , z. song , x.f . liu , p. zanardi , c.p . sun , phys . lett . , * 96 * , 140604 ( 2006 ) f.m . cucchietti , s. fernandez - vidal , j.p . paz , arxiv : quant - ph/0604136 a.e . allahverdyan , r. balian and t.m . nieuwenhuizen , phys . a , * 64 * , 032108 ( 2001 ) a. osterloh , l. amico , et al . nature ( london ) * 416 * , 608 ( 2002 ) s. hill , w.k . wootters , phys . * 78 * , 5022 ( 1997 ) w.k . wootters , phys . lett . * 80 * , 2245 ( 1998 ) s.l . zhu , phys . lett . , * 96 * , 077206 ( 2006 ) m.m . wolf , g. ortiz , f. verstraete , j.i . cirac , arxiv : cond - mat/0512180 a. langari , phys . b , * 69 * , 100402 ( 2004 ) p.w . anderson , phys . , * 86 * , 694 ( 1952 ) n.m . hugenholtz and d. pines , phys . , * 116 * , 489 ( 1959 ) r. botet and r. jullien , phys . b , * 27 * , 613 ( 1983 ) y. aharonov and a. casher , phys . rev . a , * 19 * , 2461 ( 1979 ) c.k . majumdar , j. phys . c , * 3 * , 911 ( 1970 ) t. tonegawa and i. harada , j. phys . jpn , * 56 * , 2153 ( 1987 ) h.t . wang , phys . b , * 65 * , 024426 ( 2002 ) l.g . marland and d.d . betts , phys . , * 43 * , 1618 ( 1979 ) c. tannous , phys . rev . b , * 32 * , 4770 ( 1985 ) b.m . mccoy , e. barouch and d.b . abraham , phys . a , * 4 * , 2331 ( 1971 ) t. giamarchi and h.j . schulz , phys . b , * 39 * , 4620 ( 1989 ) d.p . landau and j. thomchick , * 50 * , 1822 ( 1979 ) [ fig : ssaf4 - 10 ] : the nearest - neighbor correlation @xmath82 in different size of the spin ring as function of coupling strength @xmath15 , parameters are @xmath54 and @xmath83 . + fig . [ fig : cor6jr0 ] : the nearest - neighbor correlation @xmath82 under three different cases of external field : @xmath41 , @xmath84 and @xmath85 , the parameter @xmath54 . + fig . [ fig : cor6b ] : the dependence of nearest - neighbor correlation @xmath82 on the applied field @xmath9 with different anisotropic parameters @xmath16 . all the results are calculated with @xmath45 . [ fig : con6jr0 ] : the comparison of concurrence for two spins in the isotropic case , @xmath54 , in three different applied field : @xmath41 , @xmath84 and @xmath85 . + fig . [ fig : con6b ] : the dependence of concurrence concurrence for two spins on the applied field @xmath9 with different anisotropic parameters @xmath16 . all the results are calculated with @xmath45 . [ fig : esbz ] : the energy levels as function of @xmath9 for given coupling strength , the system size is @xmath46 , some of the levels which do not take part in the level crossing were omitted for clarity . the other parameters are : @xmath86 , @xmath45 .
the ground state of a one - dimensional spin-1/2 chain with periodical boundary condition in the heisenberg xy model is investigated . we consider the spatial correlation and concurrence between any nearest - neighbor pair of spins under the conditions of different coupling strength , anisotropic parameter and magnitude of a transverse field . quantum phase transitions due to the competition between coupling and alignment which cause the abrupt changes of the correlation and concurrence are observed . the transition are direct results of the level crossing .
let @xmath17 and let @xmath18 be a suitable bounded domain in @xmath2 . we consider the nonlinear problem @xmath19 where @xmath20 stands for the @xmath9-hessian operator of @xmath21 and @xmath22 is a given nonlinear source . problem has been studied extensively by many authors in different settings . see e.g. @xcite . + the @xmath9-hessian operator @xmath23 is defined as follows . let @xmath24 , @xmath25 , and let @xmath26 be the eigenvalues of the hessian matrix @xmath27 . then the @xmath9-hessian operator is given by @xmath28 where @xmath29 is the @xmath9-th elementary symmetric polynomial in the eigenvalues @xmath30 , see e.g. @xcite . note that @xmath31 is a family of operators which contains the laplace operator ( @xmath13 ) and the monge - ampre operator ( @xmath32 ) . the monograph @xcite is devoted to applications of monge - ampre equations to geometry and optimization theory . this family of operators has been studied extensively , see e.g. @xcite and the references therein . recently , this class of operators has attracted renewed interest , see e.g. @xcite . + we point out that the @xmath9-hessian operators are fully nonlinear for @xmath33 . further , they are not elliptic in general , unless they are restricted to the class @xmath34 observe that @xmath35 belongs to the class of subharmonic functions . further , the functions in @xmath35 are negative in @xmath18 by the maximum principle , see @xcite . the @xmath9-hessian operator defined on @xmath35 imposes certain geometry restrictions on @xmath18 . more precisely , domains called admissible are those whose boundary @xmath36 satisfies the inequality @xmath37 where @xmath38 denote the principal curvatures of @xmath36 relative to the interior normal . a typical example of a domain @xmath18 for which holds is a ball . for more details we refer the interested reader to @xcite . problem can be easily reformulated in order to study positive solutions under the change of variable @xmath39 , which in turn yields @xmath40 by the @xmath9-homogeneity of the @xmath9-hessian operator . now observe that , if @xmath41 , then the right hand side of ( [ eq : f ] ) must be nonnegative . typical examples of nonlinear terms @xmath22 appearing in the literature are @xmath42 ( see @xcite ) , @xmath43 ( see @xcite for @xmath13 and @xcite for @xmath44 ) and @xmath45 ( see @xcite for @xmath13 . ) the seminal contribution on the analysis of critical values for @xmath13 with a polynomial and exponential source was made by joseph and lundgren in @xcite . in general , for problems of gelfand type for @xmath46 , the first result ( @xmath32 ) in the radial case is due to clment et al . @xcite and for @xmath44 to jacobsen @xcite . next , for our purposes we give some general notions of solutions to ( [ eq : f ] ) . as usual , a classical solution ( or solution ) of ( [ eq : f ] ) is a function @xmath41 satisfying the equation in ( [ eq : f ] ) . we recall the version of the method of super and subsolutions for , see ( * ? ? ? * theorem 3.3 ) for more details . a function @xmath47 is called a _ subsolution _ _ supersolution _ ) of if @xmath48 note that the trivial function @xmath49 is always a supersolution . the following concept is needed to establish a general result on the existence of solutions to problem . we say that a function @xmath50 is a _ maximal _ solution of if @xmath50 is a solution of and , for each subsolution @xmath21 of , we have @xmath51 . we note that the notion of maximal solution of seems to be new in the context of @xmath9-hessian equations . in case the @xmath13 , this notion corresponds to the usual minimal ( positive ) solution , see e.g. the monograph @xcite and the references therein . in this article we study problem ( [ eq : f ] ) on the unit ball @xmath1 of @xmath2 with a polynomial source , i.e , the problem @xmath52 where @xmath53 is a parameter and @xmath54 . we recall that the @xmath9-hessian operator in radial coordinates can be written as @xmath55 , where @xmath56 is defined by @xmath57 and @xmath58 denotes the binomial coefficient . next , in order to state our main result , we write in radial coordinates , i.e. , @xmath59 now we introduce the space of functions @xmath60 defined on @xmath61 as in , for problem @xmath62 : @xmath63 ) : \left(r^{n - i}(u')^i \right)'\geq 0\;\;\mbox{in}\;\;(0,1 ) , \ , i=1, ... ,k,\,u'(0)=u(1)=0\}.\ ] ] we note that the functions in @xmath60 are non positive on @xmath64 $ ] . however , if @xmath65 for all @xmath66 , then any function in @xmath60 is negative and strictly increasing on @xmath67 . this in turn implies that , if we are looking for solutions of ( @xmath68 ) in @xmath60 , then the parameter @xmath69 must be positive . let @xmath5 . we say that a function @xmath70)$ ] is : * a _ classical solution _ of ( @xmath68 ) if @xmath71 and the equation in ( @xmath68 ) holds ; * an _ integral solution _ of ( @xmath68 ) if @xmath21 is absolutely continuous on @xmath72 $ ] , @xmath73 , @xmath74 and the equality @xmath75 holds whenever the integral exists . the concept of integral solution was introduced in @xcite for a more general class of radial operators , see e.g. @xcite and the references therein . the standard concept of weak solution is equivalent in this case to the notion of integral solution , see ( * ? ? ? * proposition 2.1 ) . the main goal of this paper is to describe the set of negative bounded radially symmetric solutions to in terms of the parameters . our statements contain some classical results ( i.e. @xmath13 ) , see @xcite . we compute a critical exponent of the joseph - lundgren type , defined by @xmath76}}{(k+1)n-2k(k+3)-2\sqrt{2[(k+1)n-2k ] } } , & n>2k+8,\\ \infty , & 2k < n \leq 2k+8 . \end{cases}\ ] ] the joseph - lundgren exponent , i.e. , @xmath77 was introduced in @xcite . we prove that @xmath14 plays the same role as the joseph - lundgren exponent . another important exponent appearing in the analysis of the boundedness of solutions to is given by @xmath78 which is smaller than @xmath14 . the value @xmath79 is well - known as the critical exponent in the study of the quasilinear @xmath9-hessian operator , see @xcite for more details . as soon as this critical exponent is crossed , a drastic change in the number of solutions of occurs . assuming that admits a solution @xmath80 for some @xmath69 , we may define the positive constant @xmath81 we show that @xmath82 is finite in theorem [ main:3:intro ] below . now we state our main result . [ main:1:thm1 ] let @xmath83 and @xmath3 . let @xmath79 and @xmath14 be as in and , respectively . * if @xmath84 and @xmath69 is close to but not equal to @xmath85 where @xmath86 , then @xmath87 has a large ( finite ) number of solutions . in addition , if @xmath88 then there exists infinitely many solutions of @xmath87 . * if @xmath89 , @xmath12 and @xmath90 , then there exists only one solution of @xmath87 . moreover , @xmath91 . for the reader s convenience , we sketch the proof of theorem [ main:1:thm1 ] . to this end , we briefly discuss , for our case , the method introduced in @xcite . we first make a rescaling of problem ( @xmath92 ) to obtain @xmath93 note that @xmath94 if and only if @xmath95 . next , we introduce a dynamical system associated to , through the following change of variables of emden - fowler type : @xmath96 that is , @xmath97 where @xmath98 the linearization of at the critical points allows for the analysis of the dynamical system on the phase plane , which depends on the eigenvalues of the jacobian matrix . the location of these eigenvalues on the complex plane is determined by the sign of @xmath99 . the case @xmath100 corresponds to instability , @xmath101 to stability , and for @xmath102 to a center . we point out that @xmath102 is equivalent to defining @xmath103 . we first note that @xmath101 is equivalent to @xmath104 and by @xcite there exists a classical global solution of . in case @xmath100 ( i.e. @xmath105 ) we can not claim the existence of a global solution of and thus discussion of this case is excluded . in the stability case , that is @xmath104 , depending on the behavior of the discriminant @xmath106 of the jacobian matrix , which is given by @xmath107,\ ] ] we obtain two types of orbits ( see figure 1 below ) . to compute the exponent @xmath14 , we first solve the equation @xmath108 and then choose the larger root of it . replacing this root into the definition of @xmath109 in , we obtain @xmath110 where @xmath111 observe that , for @xmath13 , @xmath112 coincides with the function @xmath22 introduced in @xcite . now we point out that the larger root of is exactly @xmath14 defined in in the case @xmath89 . [ [ section ] ] -@xmath113 . ] in figure 1 , the graph of @xmath114 shows that we have two possibilities : ( @xmath109 ) @xmath115 if @xmath12 or ( @xmath116 ) @xmath117 if @xmath118 . further , we see that @xmath114 decays asymptotically to the horizontal line @xmath119 , which ensures the existence of a unique positive number @xmath14 whenever @xmath120 , by . on the other hand , we see in the phase plane two heteroclinic trajectories connecting the critical points @xmath121 to @xmath122 of system . the homoclinic trajectory corresponds to the usual critical exponent @xmath123 which is analysed in our second result , see theorem [ main:2:intro ] below . finally , the number of solutions to ( @xmath68 ) is in exact correspondence with the number of intersections between a suitable horizontal line and the orbits ( a spiral or stable node ) of system . from this , we can construct a unique , large finite or infinite number of solutions to ( @xmath68 ) which is achieved following basically the arguments in @xcite . we now consider the critical exponent problem @xmath124 with @xmath79 as in . since we are looking for radially symmetric solutions to , we introduce a family of bliss functions @xmath125^{\frac{n-2k}{2(k+1)}}}{\left ( 1 + d|x|^2\right)^{\frac{n-2k}{2k}}},\;\ ; x\in { \mathbb{r}}^n,\ ] ] for @xmath126 ( see @xcite and @xcite ) and then we write the solutions of in terms of the functions @xmath127 as in . for this , note that @xmath128 solves the problem @xmath129 for all @xmath126 . restricting the functions @xmath127 to the unit ball , we define : @xmath130 where @xmath131 are the only two solutions of @xmath132 provided that @xmath69 is less than or equal to @xmath133 . when the parameter @xmath69 is equal to this last value , we have only one solution @xmath134 . now , we can state our second result . [ main:2:intro ] let @xmath3 . consider the functions @xmath135 , @xmath136 , and @xmath137 as in . set @xmath138 * if @xmath139 , then there exist exactly two solutions of given by @xmath140 * if @xmath141 , then has a unique solution given by @xmath142 we note that the value @xmath143 , hereafter denoted by @xmath144 , is computed by the rescaling @xmath145 . by applying the boundary condition on @xmath50 , and noting that @xmath50 is a bliss function by ( * ? ? ? * lemma 5.2 ) , we obtain @xmath144 . notice that , for @xmath123 , we have the estimate @xmath146 by theorem [ main:3:intro ] ( see below ) . we mention that , in the case @xmath13 , the value @xmath147 coincides with the extremal value obtained in the classical paper @xcite . see also @xcite and @xcite . we point out that , if any solution of the problem @xmath148 is radially symmetric around the origin , then we have @xmath149 . the study of radially symmetric solutions to in @xmath2 seems not to have been discussed in the literature , except if we restrict to a ball , see @xcite . however , for the laplacian operator ( @xmath13 ) and for the @xmath150-laplacian operator , there exist well - known references about radially symmetric solutions to the corresponding problems , see e.g. @xcite . our third result deals with the existence of a finite positive number @xmath151 , as in , such that problem @xmath62 has either negative maximal bounded solutions if @xmath152 , at least one integral ( or weak ) solution ( possibly unbounded ) when @xmath153 , or no classical solutions when @xmath154 . more precisely , [ main:3:intro ] let @xmath155 and @xmath6 . then there exists @xmath156 such that : * if @xmath157 , then @xmath62 admits a maximal bounded solution . * if @xmath153 , then @xmath62 admits at least one possible unbounded integral solution . * if @xmath158 , then @xmath62 admits no classical solutions . note that in the case @xmath13 , i.e. , for the laplace operator , it is well - known that there exists a finite positive ( extremal ) parameter @xmath82 such that problem ( [ eq : f : pol ] ) has a positive minimal classical solution @xmath159 if @xmath160 , while no solution exists , even in the weak sense , for @xmath161 . ( see @xcite and the references therein ) . the boundedness of solutions for the extremal value @xmath151 , which are so called _ extremal solutions _ , depends on the dimension @xmath162 and on critical values of the power @xmath7 . in @xcite the aim was to study the properties of the extremal solutions of problem ( [ eq : f : pol ] ) . further , in this reference the main interest centered on the unbounded solutions ( i.e. singular solutions or blow - up solutions ) . for @xmath163 and @xmath164 we have the explicit weak solution @xmath165 corresponding to the parameter @xmath166 thus problem ( [ eq : f ] ) admits at least one solution if @xmath167 ( see@xcite ) . in our case , for @xmath3 and @xmath168 , we introduce the radial function @xmath169 defined by @xmath170 this function is an explicit integral solution of problem ( [ eq : f : pol ] for more details ) , corresponding to the parameter @xmath171 provided that @xmath12 holds . note that , for @xmath13 , we have @xmath172 and @xmath173 . to prove the statement ( a ) we need two lemmas . the first lemma is a variant of ( * ? ? ? * lemma 4 ) tailored to our needs . [ bcmr96 ] let @xmath174 be a @xmath175 convex and nondecreasing function on @xmath176 . assume that @xmath177 and set @xmath178 let @xmath179 be a @xmath175 positive function on @xmath176 such that @xmath180 and @xmath181 . now set @xmath182 and @xmath183 for all @xmath184 . then clearly @xmath192 . further , since @xmath193 and @xmath194 are decreasing functions , we have @xmath186 for all @xmath184 . thus ( i ) holds . property ( ii ) is clear . to prove ( iii ) , we have @xmath195 and @xmath196 since @xmath197 it follows that @xmath190 is a convex function on @xmath198 $ ] , which completes the proof . fix @xmath202 and define the functions @xmath203^{1/k } \text { and } \tilde{g}(t ) = \left[\lambda(1+t)^q\right]^{1/k},\text { for all } t\geq 0.\ ] ] set @xmath183 ( @xmath184 ) with @xmath193 and @xmath194 as in lemma [ bcmr96 ] . since @xmath83 , we have that @xmath187 exists and hence @xmath190 is bounded by lemma [ bcmr96 ] ( ii ) . next , by and lemma [ bcmr96 ] we have @xmath204 therefore , @xmath205 is a bounded subsolution of @xmath87 and hence by the method of super and subsolutions we have , by ( * ? ? ? * theorem 3.3 ) , a solution @xmath206 of @xmath87 with @xmath207 . now , to prove that ( @xmath92 ) admits a maximal solution we consider @xmath208 as a solution of @xmath209 since @xmath21 is in particular a subsolution of ( @xmath92 ) , we have @xmath210 on @xmath1 by the comparison principle , see @xcite . next , we define @xmath211 ( @xmath212 ) as a solution of @xmath213 using again the comparison principle we obtain a decreasing sequence of @xmath211 bounded from below by @xmath21 and by 0 from above . hence , we can pass to the limit and we obtain a solution @xmath214 of ( @xmath92 ) , which is maximal since the recursive sequence @xmath215 does not depend on the subsolution @xmath21 . now , let @xmath216 and @xmath217 , @xmath218 be maximal solutions of @xmath219 ( @xmath220 ) , respectively . note that @xmath218 is a subsolution of @xmath221 , whence @xmath222 by the maximality of @xmath217 . now , we show that @xmath151 finite and positive . for @xmath223 , let @xmath224 be a ball centered at zero with radius @xmath225 such that @xmath226 and let @xmath227 be the solution of @xmath228 then there exists a negative constant @xmath229 such that @xmath230 on @xmath231 . set @xmath232 and take @xmath233 . then @xmath234 by ( * ? ? ? * theorem 3.3 ) , for any @xmath235 there exists a solution @xmath80 of @xmath87 . hence @xmath156 . to see that @xmath151 is finite we consider the inequality @xmath236^\frac{1}{k},\ ; u\in \phi^k(b).\ ] ] see e.g. let @xmath237 be the first eigenvalue of @xmath238 with zero dirichlet boundary condition . let @xmath152 and let @xmath21 be a solution of problem @xmath87 . then , using , we obtain @xmath239^\frac{1}{k}=c(n , k)\lambda^{\frac{1}{k}}(1-u)(1-u)^{\frac{q - k}{k}}>c(n , k)\lambda^{\frac{1}{k}}(1-u)>\lambda^{\frac{1}{k}}c(n , k)(-u),\ ] ] which in turn implies @xmath240 . thus @xmath241 is finite . now let @xmath245 be an increasing sequence such that @xmath246 as @xmath247 and let @xmath248 be a maximal solution of @xmath219 . by lemma [ max : sol ] , for all @xmath249 $ ] , we have @xmath250 . on the other hand , integrating the equation in @xmath219 , we obtain @xmath251^{\frac{1}{k}}d\tau.\ ] ] now , applying twice the monotone convergence theorem we conclude that @xmath252 and @xmath253^{\frac{1}{k}}d\tau , \;\;\ ; \text { a.a . } r\in ( 0,1).\ ] ] this proves assertion ( b ) . the rescaled function @xmath254 solves the problem @xmath255 in particular , in radial coordinates @xmath256 , the indicates function solves the ordinary differential equation @xmath257 with the boundary conditions @xmath258 and @xmath259 . now define @xmath260 , with @xmath261 , as the maximal continuation interval for @xmath262 under the restriction @xmath263 further , after integration of over @xmath264 , with @xmath265 , we obtain @xmath266 we claim that @xmath267 . suppose by contradiction that @xmath268 . since @xmath50 is increasing by and bounded from above by @xmath269 , we have @xmath270 by the maximality of @xmath225 . therefore @xmath50 is a nontrivial solution of @xmath271 which contradicts tso s nonexistence result , see ( * ? ? ? * proposition 1 ) . next , after scaling @xmath272 we can rewrite as @xmath273 choosing @xmath274 , @xmath275 , and @xmath276 in ( * ? ? ? * equation ( 1.12 ) ) we conclude that admits a unique solution of the form @xmath277^{\frac{n-2k}{2(k+1)}}}{\left ( 1 + d|x|^2\right)^{\frac{n-2k}{2k}}},\,\ , x\in { \mathbb{r}}^n,\ ] ] for @xmath126 ( see @xcite and @xcite ) . now we can write the solutions of in terms of the functions @xmath127 as . more precisely , it is easy to see that @xmath127 is a solution of if , and only if , the there exists a value @xmath126 such that @xmath278 we can now verify by elementary calculus that has either a unique solution @xmath279 if @xmath280 , exactly two solutions @xmath281 if @xmath282 and no solutions @xmath283 if @xmath284 . hence problem has a solution if , and only if , @xmath285 . therefore , by the rescaled function @xmath50 , we conclude that problem has exactly two solutions @xmath243 , @xmath289 if @xmath286 and a unique solution @xmath290 if @xmath291 , where @xmath292 and @xmath293 the proof is now complete . let @xmath3 and @xmath104 . set @xmath294 and introduce the nonnegative variable @xmath295 with @xmath296 as in . hereafter @xmath296 is denoted by @xmath297 . note that @xmath298 since @xmath299 . define the rescaled function @xmath300 as @xmath301 we note that , by @xmath62 and , @xmath50 solves the initial value problem @xmath302 let @xmath305 and set @xmath306 . then we may rewrite as @xmath307 defining @xmath308 , @xmath309 , with @xmath274 , @xmath310 , @xmath276 , @xmath311 and @xmath312/(\beta + 1)$ ] , we obtain @xmath313 for @xmath309 if , and only if , @xmath314 . then the global existence of follows from ( * ? ? ? * theorem 4.1 ) . the uniqueness follows from the contraction mapping theorem as we show next . we use the notation given in ( * ? ? ? * theorem 4.1 ) and define the map @xmath315 , where @xmath316\;:\ ; w(0)=-1 \text { and } |w+1| \leq 1/2\}$ ] , by @xmath317.\ ] ] using the arguments given in the proof of ( * ? ? ? * theorem 4.1 ) , we see that the map @xmath318 satisfies @xmath319 and admits a fixed point . we show now that @xmath318 is a contraction . to this end , let @xmath320 be fixed , and define @xmath321 it is easy to see that @xmath322 on the other hand , from the inequality @xmath323 we conclude that there exists a constant @xmath324 such that @xmath325 hence we have the estimate @xmath326}|w_1(s)-w_2(s)|,\end{aligned}\ ] ] since @xmath327 for some positive constant @xmath328 and for all @xmath329 $ ] . we choose @xmath330 such that @xmath331 , which ensures the existence of a unique local solution . using standard arguments of continuation on the maximal existence interval , we deduce the existence of a unique global solution on @xmath176 . next , we introduce a dynamical system associated to by means of the following change of variables of emden - fowler type : @xmath332 and the introduction of the functions @xmath333 after a straightforward ( and lengthy ) computation , we conclude that the pair @xmath334 satisfies the following system of differential equations @xmath335 where @xmath109 is defined as in . moreover , the function @xmath336 satisfies @xmath337 by , and . this in turn implies that @xmath113 vanishes as @xmath338 . analogously , we can obtain , by , and that @xmath339 for @xmath314 , the unique global solution of corresponds , by lemma [ uniq : sol ] , to a global solution @xmath340 of - . on the other hand , a global solution @xmath340 of - defines , by lemma [ uniq : sol ] and reversing the definitions of - , a unique solution @xmath50 of . thus we have a one - to - one correspondence between solutions of and solutions of - . we have already shown the necessity of the condition by - . now assume that @xmath342 from the analysis of the field directions to we see that , for @xmath343 small enough , @xmath344 and @xmath345 , which may be used to estimate @xmath346 and @xmath347 . to this end , we combine the inequalities arising from @xmath344 and @xmath345 and obtain @xmath348 where @xmath349 is chosen small enough . next , by , we may integrate over @xmath350 the quotient @xmath351 after the change of variable @xmath352 , we obtain the estimate @xmath353 thus , by definition of @xmath113 in , we have @xmath354 since @xmath50 is increasing by the equation in , we obtain that @xmath50 is globally bounded . the initial condition @xmath355 follows by lhospital s rule and the equality @xmath356 follows from . the dynamical system has exactly two stationary points in the phase plane @xmath340 . we denote these two points by @xmath121 and @xmath122 , where @xmath357 now we consider the linearization at the equilibrium point @xmath122 of . the jacobian matrix at @xmath122 is given by @xmath358 the eigenvalues of @xmath359 are @xmath360 where @xmath361 and @xmath362 . the discriminant is given by @xmath107.\ ] ] the location on the complex plane of the eigenvalues @xmath363 is determined as follows : * if @xmath100 and @xmath364 , then the eigenvalues @xmath363 are real positive numbers . * if @xmath100 and @xmath365 , then the eigenvalues @xmath363 are complex numbers with positive real part . * if @xmath101 and @xmath365 , then the eigenvalues @xmath363 are complex numbers with negative real part . * if @xmath101 and @xmath364 , then the eigenvalues @xmath363 are negative real numbers . * if @xmath102 , then the eigenvalues @xmath363 are purely imaginary . we observe that the cases ( i ) and ( ii ) can be ignored since @xmath100 is equivalent to @xmath366 . the condition @xmath102 in ( v ) is equivalent to setting @xmath367 , and this case was already discussed in theorem [ main:2:intro ] . additionally , we note that the orbit @xmath368 of - in the critical case @xmath367 convergences to @xmath121 after one loop around @xmath122 . indeed , let @xmath50 the unique solution of . then , using tso s nonexistence result and the arguments in the proof of theorem [ main:2:intro ] , we conclude that @xmath50 solves the problem @xmath369 then , after rescaling @xmath370 , we see that @xmath371 solves the problem for some @xmath283 . the initial condition @xmath356 yields the existence of three values of @xmath283 as in the proof of theorem [ main:2:intro ] . since @xmath371 is a bliss function , we can compute explicitly @xmath50 , which is given by @xmath372 hence , by the change of variable and , we have @xmath373^ke^{\frac{(n+2)k}{k+1}t}\left(1+de^{2t}\right)^{-\frac{n}{2}}\\ z(t ) & = \tilde{\lambda}e^{\frac{(n+2)k}{k+1}t } \left(1+de^{2t}\right)^{-\frac{n+2}{2}}.\end{aligned}\ ] ] therefore , @xmath374 since @xmath3 . let @xmath376 and @xmath377 be as in . then , by , we have @xmath378 on the other hand , by , , and lhospital s rule , we deduce that @xmath379 which in turn implies that @xmath380 thus the lemma follows immediately by combining and . we first describe the behavior of the orbit @xmath368 near @xmath122 . to this end , we compute @xmath384 by lhospital s rule . the existence of the limit above is equivalent to solving the equation @xmath385 whose roots are given by @xmath386 note that @xmath387 if , and only if , @xmath12 . thus , @xmath388 are positive roots . next , consider a function @xmath174 defined by @xmath389 where the constants @xmath390 and @xmath229 are chosen such that the graph of @xmath174 connects the points @xmath121 and @xmath122 . using @xmath391 , we obtain @xmath392 . now , setting @xmath393 , we have @xmath394 . note that @xmath395 since @xmath12 and @xmath396 . we claim that the unique solution @xmath334 lies below @xmath397 when @xmath398 . indeed , by lemma [ asym : z:1 ] , @xmath334 lies above the line @xmath399 and below @xmath397 near @xmath121 on the phase plane . on the other hand , since the slope of the line @xmath399 at @xmath122 is larger than @xmath400 and the parabola in is positive at @xmath401 , we conclude that the orbit arrives at @xmath122 from above the line @xmath399 near @xmath122 . suppose now by contradiction that @xmath334 intersects the curve @xmath397 in a point @xmath402 at @xmath403 . in this case we have two possibilities : the orbit @xmath334 remains on the graph of @xmath174 for all @xmath404 and then @xmath334 arrives at @xmath122 with the same slope that @xmath397 at @xmath405 , which is impossible since @xmath406 because the inequality @xmath407 . the other case is that the orbit crosses the graph of @xmath174 at a point @xmath408 . in this case we have @xmath409 we now show that is impossible . to this end , define the functions @xmath410 for all @xmath411\times [ 0,z_2]$ ] . note that @xmath412 next , set @xmath413 then , by and , we have @xmath414 using , we obtain @xmath415 ^ 2 + \frac{2k+a}{q - k}\,g'(y_2 ) - \frac{2aq}{(q - k)^2}\geq 0\ ] ] since @xmath12 and the inequality is strict if @xmath416 . now note that @xmath417 since @xmath418 and @xmath419 by and the equality @xmath391 . the derivative of @xmath193 may be written as @xmath420 where the function @xmath421 is given by @xmath422y^{1/k - \beta / q } -\beta ( 2\beta-1)c^2y^{\beta-1}.\ ] ] now , in order to determine the growth of @xmath193 around @xmath269 , we rewrite @xmath423 as @xmath424 and conclude that @xmath425 since @xmath426 and @xmath418 . note that @xmath427 by - . in particular , @xmath428 in the case @xmath429 by and the equalities @xmath430 . on the other hand , it is easy to see that @xmath431 computing , @xmath432 we conclude that the equation @xmath433 admits only one solution which is not a minimum by . further , the equation @xmath434 has at most two solutions on @xmath435 $ ] since , if there is no solution , then @xmath436 on @xmath435 $ ] by . this would in turn imply that @xmath193 is decreasing on @xmath435 $ ] by , which is a contradiction since @xmath437 . now , if we have two solutions on @xmath438 then @xmath439 , which is a contradiction since @xmath427 . hence the equation @xmath434 admits at most two solutions on @xmath435 $ ] . in this case we see that @xmath193 admits at most two critical points on @xmath435 $ ] . but this is impossible since @xmath440 by - and and . the proof is now complete . [ qk : espiral ] let @xmath441 . let @xmath368 be the unique solution of - and @xmath50 the unique solution of . then @xmath368 loops around @xmath122 an infinite number of times and converges to @xmath122 as @xmath442 . we have the following asymptotic behavior for @xmath50 at infinity @xmath443 we first show that the trajectory does not converge to @xmath121 as @xmath442 . suppose by contradiction that @xmath368 converges to @xmath121 as @xmath442 . then , for all @xmath343 large enough , @xmath346 and @xmath347 are decreasing functions . hence , from we have @xmath444 where @xmath445 is large enough and @xmath446 is a positive constant . since @xmath447 are two vanishing functions as @xmath442 and @xmath83 , by ( [ eq : dyn ] ) and ( [ boundzeta ] ) , we conclude that @xmath448 using ( [ asympyprima ] ) , we deduce that , for any @xmath449 , there exists @xmath450 such that @xmath451 and this , together with ( [ y - z : def ] ) and ( [ boundzeta ] ) , yields @xmath452 where @xmath453 is a constant depending on @xmath454 . choosing @xmath454 small enough in ( [ bound - v ] ) , we see that the function @xmath455 belongs to @xmath456 and then by ( [ eq : v ] ) we obtain @xmath457 in turn now by ( [ simvprima ] ) and lhospital s rule we obtain @xmath458 next , choosing @xmath459 and @xmath460 in ( * ? ? ? * proposition 4.1 ) , we conclude that , for each @xmath461 , the integrals on the left hand side in ( * ? ? ? * proposition 4.1 ) are zero . consequently @xmath462 using - , we deduce that the first two terms on the right hand side of ( [ identitypps ] ) go to zero as @xmath463 since @xmath3 , and the last term vanishes as @xmath463 since @xmath104 , which in turn yields @xmath464 this is a contradiction . therefore , the orbit @xmath368 converges to the stationary point @xmath122 or to a limit cycle . the existence of a limit cycle is excluded by the generalized version of bendixson s theorem . indeed , define @xmath465 then @xmath466 since @xmath104 . thus there is no limit cycle of by the bendixson - dulac theorem . hence @xmath368 converges to @xmath122 as @xmath442 after infinitely many loops around @xmath122 since the eigenvalues of the linearized system in @xmath122 are complex numbers in view of @xmath118 . the asymptotic estimate ( [ asymptotic ] ) is an immediate consequence of the change of variables - . next , returning to the proof of the theorem , we show the uniqueness and the multiplicity of solutions of @xmath62 as follows : let @xmath467 be fixed , define @xmath468 and set @xmath469 . since the component @xmath113 of the orbit satisfies @xmath470 as @xmath442 for all @xmath104 by lemmas [ lem : qjl : nodo ] and [ qk : espiral ] , we have @xmath471 as @xmath472 and @xmath473 as @xmath472 by the change of variable - . now , using the changes of variables - together with , we obtain the line @xmath474 note that this line has range @xmath475 . further @xmath476 and @xmath477 by . moreover , for each intersection between the line and the orbit , we obtain one and/or several times @xmath478 . for each @xmath478 we define @xmath479 with @xmath480 and @xmath481 is given by . hence , in the case @xmath482 , as time increases , the line intersects the orbit @xmath368 either a finite large number of times when @xmath69 approaches @xmath297 , or an infinite number of times when @xmath483 . in the case @xmath12 , we see that the line intersects the orbit @xmath368 only one time for each time @xmath478 , which in turn implies that the problem @xmath484 admits a unique solution @xmath485 . now , if @xmath486 we see that the solution @xmath485 is a maximal solution by theorem [ main:3:intro ] , which is decreasing as @xmath487 increases . since @xmath488 as @xmath489 , we conclude that @xmath490 as @xmath489 since @xmath486 , which is impossible by the change of variable - . therefore , @xmath491 . this completes the proof . f. brock . radial symmetry for nonnegative solutions of semilinear elliptic equations involving the @xmath150-laplacian . in _ progress in partial differential equations , vol . 1 ( pont - - mousson , 1997 ) _ , volume 383 of _ pitman res . notes math . ser . _ , pages 4657 . longman , harlow , 1998 . b. gidas , w .- m . ni , and l. nirenberg . symmetry of positive solutions of nonlinear elliptic equations in @xmath496 . in _ mathematical analysis and applications , part a _ , volume 7 of _ adv . in math . _ , pages 369402 . academic press , new york - london , 1981 .
we consider the problem @xmath0 where @xmath1 denotes the unit ball in @xmath2 , @xmath3 ( @xmath4 ) , @xmath5 and @xmath6 . we study the existence of negative bounded radially symmetric solutions of . in the critical case , that is when @xmath7 equals tso s critical exponent @xmath8 , we obtain exactly either one or two solutions depending on the parameters . further , we express such solutions explicitly in terms of bliss functions . the supercritical case is analysed following the ideas develop by joseph and lundgren in their classical work @xcite . in particular , we establish an emden - fowler transformation which seems to be new in the context of the @xmath9-hessian operator . we also find a critical exponent , defined by @xmath10}}{(k+1)n-2k(k+3)-2\sqrt{2[(k+1)n-2k ] } } , & n>2k+8,\\ \infty , & 2k < n \leq 2k+8 , \end{cases}\ ] ] which allows us to determinate the multiplicity of the solutions to int the two cases @xmath11 and @xmath12 . moreover , we point out that , for @xmath13 , the exponent @xmath14 coincides with the classical joseph - lundgren exponent . @xmath15 + email : [email protected] @xmath16 + email : [email protected] 2010 mathematics subject classification . primary : 35b33 ; secondary : 34c37 , 34c20 , 35j62 . _ keywords and phrases . _ @xmath9-hessian operator ; radial solutions ; critical exponents ; phase analysis ; emden - fowler transformation .
the evolution of a collisionless self - gravitating system is described by two coupled equations : the vlasov equation , @xmath1 where @xmath2 is the one - particle distribution function and @xmath3 is the gravitational potential , and poisson s equation , @xmath4 @xmath0-body simulations use a monte - carlo method to solve these equations . the distribution function is represented by a collection of @xmath0 particles @xcite : @xmath5 where @xmath6 , @xmath7 , and @xmath8 are the mass , position , and velocity of particle @xmath9 . over time , particles move along characteristics of ( [ eq : vlasov ] ) ; at each instant , their positions provide the density needed for ( [ eq : poisson ] ) . in many collisionless @xmath0-body simulations , the equations of motion actually integrated are @xmath10 where @xmath11 is the _ softening length_. these equations reduce to the standard newtonian equations of motion if @xmath12 . the main reason for setting @xmath13 is to suppress the @xmath14 singularity in the newtonian potential ; this greatly simplifies the task of numerically integrating these equations ( e.g. , * ? ? ? * ) . by limiting the spatial resolution of the gravitational force , softening also helps control fluctuations caused by sampling the distribution function with finite @xmath0 ; however , this comes at a price , since the gravitational field is systematically biased for @xmath13 @xcite . softening is often described as a modification of newtonian gravity , with the @xmath14 potential replaced by @xmath15 . the latter is proportional to the potential of a @xcite sphere with scale radius @xmath11 . this does _ not _ imply that particles interact like plummer spheres @xcite ; the acceleration of particle @xmath9 is computed from the field at the point @xmath7 only . but it does imply that softening can also be described as a smoothing operation ( e.g. , * ? ? ? * ) , in which the pointillistic monte - carlo representation of the density field is convolved with the kernel @xmath16 in effect , the source term for poisson s equation ( [ eq : poisson ] ) is replaced with the _ smoothed density _ @xmath17 formally , ( [ eq : nbody ] ) provides a monte - carlo solution to the vlasov equation ( [ eq : vlasov ] ) coupled with @xmath18 thus one may argue that a softened @xmath0-body simulation actually uses standard newtonian gravity , as long as it is clear that the mass distribution generating the gravitational field is derived from the particles via a smoothing process . although plummer softening is widely used in @xmath0-body simulations , its effects are incompletely understood . if the underlying density field is featureless on scales of order @xmath11 , softening has relatively little effect . however , @xmath0-body simulations are often used to model systems with power - law density profiles ; for example , @xcite and ( * ? ? ? * hereafter nfw ) models , which have @xmath19 at small @xmath20 , are widely used as initial conditions . one purpose of this paper is to examine how softening modifies such profiles . assume that the underlying density profile is spherically symmetric and centered on the origin : @xmath21 . the integrand in ( [ eq : smooth - rho ] ) is unchanged by rotation about the axis containing the origin and the point @xmath22 , so the integral can be simplified by adopting cylindrical coordinates @xmath23 , where @xmath22 is located on the @xmath24-axis at @xmath25 . the integral over @xmath26 is trivial ; for plummer smoothing , the result is @xmath27 where the second equality holds because the outer integral is taken over the entire @xmath24 axis . the first step is to examine the effect of plummer smoothing on power - law density profiles , @xmath28 , where @xmath29 . these profiles are not realistic , since the total mass diverges as @xmath30 . however , results obtained for power - law profiles help interpret the effects of smoothing on more realistic models . profile . dashed line is the underlying density profile ; solid curve is the result of smoothing with @xmath32 . the smoothed profile slightly exceeds the underlying power - law at large @xmath20 . [ cusp2 ] ] to 0.10 profile . dashed line is the underlying density profile ; solid curve is the result of smoothing with @xmath32 . the smoothed profile slightly exceeds the underlying power - law at large @xmath20 . [ cusp2 ] ] let the density be @xmath33 . the total mass enclosed within radius @xmath20 is @xmath34 . in this case , the smoothed density profile can be calculated analytically ; the double integral is @xmath35 this yields a remarkably simple result for the smoothed density , plotted in fig . [ cusp1 ] , @xmath36 where @xmath37 . the smoothed mass within radius @xmath20 , hereafter called the _ smoothed mass profile _ ; only density profiles can be smoothed . ] , is @xmath38 let the density be @xmath40 . the total mass enclosed within radius @xmath20 is @xmath41 . the integral over @xmath42 can be evaluated , but the result is not particularly informative and the remaining integral must be done numerically . [ cusp2 ] presents the results . for @xmath43 , the smoothed density exceeds the underlying power - law profile . this occurs because smoothing , in effect , spreads mass from @xmath44 to larger radii , and with the underlying profile dropping away so steeply this redistributed mass makes a relatively large contribution to @xmath45 . note that as @xmath46 , the smoothed density @xmath47 . it appears impossible to calculate the smoothed density profile for arbitrary @xmath48 without resorting to numerical methods , but the central density is another matter . setting @xmath49 , the smoothed density is @xmath50 the central density ratio @xmath51 is plotted as a function of @xmath48 in fig . [ dzero ] . for @xmath52 and @xmath53 , the ratio @xmath54 and @xmath53 , respectively , in accord with the results above , while as @xmath55 the central density diverges . the smoothed central density for an arbitrary power - law is useful in devising an approximate expression for the smoothed density profile ( appendix a.1 ) . in addition , the central density is related to the shortest dynamical time - scale present in an @xmath0-body simulation , which may in turn be used to estimate a maximum permissible value for the time - step ( 4.3.1 ) . plotted as a function of @xmath48 . limiting values are @xmath54 as @xmath56 and @xmath57 as @xmath55 . [ dzero ] ] as noted above , both of these profiles have @xmath31 as @xmath46 . for this reason , they are treated in parallel . the @xcite model has density and mass profiles @xmath58 where @xmath59 is the scale radius and @xmath60 is the total mass . the @xcite model has density and mass profiles @xmath61 where @xmath59 is again the scale radius and @xmath62 is a characteristic density . the double integrals required to evaluate the smoothed versions of these profiles appear intractable analytically but can readily be calculated numerically . [ rhoh ] and [ rhonfw ] present results for a range of @xmath11 values between @xmath59 and @xmath63 . for comparison , both models are scaled to have the same underlying density profile at @xmath64 . and density @xmath65 . lower curves show profiles smoothed with @xmath66 , @xmath67 , , @xmath68 ( from top to bottom ) ; heavy curve is @xmath69 , dashed curve is @xmath32 . inset shows ratio @xmath70 . [ rhonfw ] ] to 0.10 and density @xmath65 . lower curves show profiles smoothed with @xmath66 , @xmath67 , , @xmath68 ( from top to bottom ) ; heavy curve is @xmath69 , dashed curve is @xmath32 . inset shows ratio @xmath70 . [ rhonfw ] ] the smoothed profiles shown in figs . [ rhoh ] and [ rhonfw ] are , for the most part , easily understood in terms of the results obtained for power - laws . for radii @xmath71 , smoothing transforms central cusps into constant - density cores , just as in fig . [ cusp1 ] . if the softening length @xmath11 is much less than the scale length @xmath59 , the smoothed density @xmath72 within @xmath64 is almost independent of the underlying profile at radii @xmath73 . consequently , the smoothed central density @xmath74 , echoing the result obtained for the power - law @xmath75 . in addition , the actual curves in figs . [ rhoh ] and [ rhonfw ] are shifted versions of the curves in fig . [ cusp1 ] ; this observation motivates simple approximations to @xmath76 and @xmath77 described in appendix a.2 . on the other hand , if @xmath11 is comparable to @xmath59 , the quantitative agreement between these profiles and the smoothed @xmath52 profile breaks down ; the smoothed density at small @xmath20 has a non - negligible contribution from the underlying profile beyond the scale radius @xmath59 . as an example , for @xmath32 the central density of the smoothed nfw profile is higher than the central density of the smoothed hernquist profile , because the former receives a larger contribution from mass beyond the scale radius . a somewhat more subtle result , shown in the insets , is that heavily smoothed profiles _ exceed _ the underlying profiles at radii @xmath78 . this is basically the same effect found with the @xmath79 power - law profile ( 2.2 ) ; with the underlying density dropping rapidly as a function of @xmath20 , the mass spread outward from smaller radii more than makes up for the mass spread to still larger radii . this effect is more evident for the hernquist profile than for the nfw profile because the former falls off more steeply for @xmath78 . the @xcite model has density and mass profiles @xmath80 where @xmath59 is the scale radius and @xmath60 is the total mass . the double integrals required to evaluate the smoothed version of this profile appear intractable analytically but can readily be calculated numerically . fig . [ rhoj ] present results for a range of @xmath11 values between @xmath59 and @xmath63 . , [ rhonfw ] , and [ rhoj ] . solid lines are underlying profiles ; from bottom to top , they represent jaffe , hernquist , and nfw models , respectively . dashed , dot - dashed , and dotted lines give results for @xmath81 , @xmath82 , and @xmath83 , respectively . [ slope ] ] to 0.10 , [ rhonfw ] , and [ rhoj ] . solid lines are underlying profiles ; from bottom to top , they represent jaffe , hernquist , and nfw models , respectively . dashed , dot - dashed , and dotted lines give results for @xmath81 , @xmath82 , and @xmath83 , respectively . [ slope ] ] again , much of the behavior shown in this plot can be understood by reference to the results for the @xmath79 power - law . in particular , for smoothing lengths @xmath84 , the central density is @xmath85 , and the curves in fig . [ rhoj ] are shifted versions of the one in fig . [ cusp2 ] . as the inset shows , for larger values of @xmath11 the smoothed profiles quite noticeably exceed the underlying profile ; the effect is stronger here than it is for a hernquist model because the jaffe model has more mass within @xmath86 to redistribute . figs . [ rhoh ] , [ rhonfw ] , and [ rhoj ] have interesting implications for @xmath0-body experiments . one might expect the smoothed profiles to resolve the inner power - laws of the underlying models as long as the softening length @xmath11 is somewhat less than the scale radius @xmath59 , but that is not what these figures show . profiles smoothed with @xmath87 are essentially constant - density cores attached to power - law outer profiles ; the density within the core depends on @xmath11 , but no inner cusp per se can be seen . for @xmath88 , on the other hand , the smoothed profiles do appear to trace the inner power - laws over some finite range of radii , before flattening out at smaller @xmath20 . only for @xmath89 can the inner cusps be followed for at least a decade in radius . [ slope ] helps explain this result . the underlying jaffe , hernquist , and nfw profiles all roll over gradually from their inner to outer power - law slopes between radii @xmath90 . thus a resolution somewhat better than @xmath91 is required to see the inner cusps of these models . in practice , this implies the softening parameter @xmath11 must be several times smaller than @xmath91 . since the formalism developed above is exact , numerical tests of a relation like ( [ eq : plummer - smooth ] ) for the smoothed density @xmath72 may seem superfluous . in practice , such tests can be illuminating as benchmarks of @xmath0-body technique . in what follows , the smoothing formalism will be applied to actual @xmath0-body calculations , to check @xmath0-body methodology and to demonstrate that the formalism has real applications . putting this plan into operation requires some care . to begin with , an @xmath0-body realization of a standard hernquist or jaffe profile spans a huge range of radii . typically , the innermost particle has radius @xmath92 or @xmath93 for a hernquist or jaffe profile , respectively , while for either profile , the outermost particle has radius @xmath94 . a dynamic range of @xmath95 or @xmath96 can be awkward to handle numerically ; even gridless tree codes may not accommodate such enormous ranges gracefully . one simple option is to truncate the particle distribution at some fairly large radius , but it s preferable to smoothly taper the density profile : @xmath97 \!\ ! ( 1 + \mu ) \ , \rho _ { * } \ , ( b / r)^2 \ , e^{-r / r _ { * } } \ , , & r > b \\ \end{array } \right . \label{eq : tapered - models}\ ] ] where the taper radius @xmath98 , the values of @xmath99 and @xmath100 are fixed by requiring that @xmath101 and its first derivative are continuous at @xmath102 , and the value of @xmath103 is chosen to preserve the total mass . let @xmath104 be the logarithmic slope of the density profile at @xmath102 , and @xmath105 be the underlying mass profile ; then @xmath106 , @xmath107 , and @xmath108 , respectively . [ rhohjt ] ] to 0.10 , @xmath107 , and @xmath108 , respectively . [ rhohjt ] ] fig . [ rhohjt ] shows how plummer smoothing modifies tapered hernquist and jaffe profiles . both profiles have scale radius @xmath109 , taper radius @xmath110 , and mass @xmath111 ; these parameters will be used in all subsequent calculations . in each case , the underlying profile follows the standard curve out to the taper radius @xmath112 , and then rapidly falls away from the outer power law . at radii @xmath113 , the smoothed profiles match those shown in figs . [ rhoh ] and [ rhoj ] , apart from the factor of @xmath114 used to preserve total mass . at larger radii , the smoothed profiles initially track the underlying tapered profiles , but then transition to asymptotic @xmath115 power law tails . this occurs because the plummer smoothing kernel ( [ eq : plummer - kernel ] ) falls off as @xmath116 at large @xmath20 ; in fact , these power laws match @xmath117 , which is the large-@xmath20 approximation for a point mass @xmath60 smoothed with a plummer kernel . the amount of mass in these @xmath116 tails is negligible . in principle , it s straightforward to verify that the smoothed profiles above generate potentials matching those obtained from @xmath0-body calculations . for a given density profile @xmath118 , construct a realization with @xmath0 particles at positions @xmath7 ; a @xmath0-body force calculation with softening @xmath11 yields the gravitational potential @xmath119 for each particle . conversely , given the smoothed density profile @xmath72 , compute the smoothed mass profile @xmath120 , and use the result to obtain the _ smoothed potential _ @xmath121 : @xmath122 with boundary condition @xmath123 as @xmath124 . for each particle , the @xmath0-body potential @xmath119 may be compared with the predicted value @xmath125 ; apart from @xmath126 fluctuations , the two should agree . a major complication is that @xmath126 fluctuations in @xmath0-body realizations imprint spatially coherent perturbations on the gravitational field ; potentials measured at adjacent positions are not statistically independent . for example , the softened potential at the origin of an @xmath0-body system is @xmath127 if the radii @xmath128 are independently chosen , this expression is a monte carlo integral , which will deviate from @xmath129 by an amount of order @xmath130 ; moreover , everywhere within @xmath131 the potential will deviate upward or downward by roughly as much as it does at @xmath49 . one way around this is to average over many @xmath0-body realizations , but this is tedious and expensive . an easier solution is to sample the radial distribution uniformly . let @xmath105 be the mass profile associated with the underlying density @xmath118 . assign all particles equal masses , and determine the radius @xmath128 of particle @xmath9 by solving @xmath132 for @xmath133 to @xmath0 . this eliminates radial fluctuations ; the monte - carlo integral for @xmath134 is replaced with a panel integration uniformly spaced in @xmath105 , and the central potential is obtained with relatively high accuracy . this trick does not suppress _ non - radial _ fluctuations , so the @xmath0-body potential evaluated at any point @xmath135 still differs from the true @xmath121 . but a non - radial fluctuation which creates an overdensity at some position @xmath136 must borrow mass from elsewhere on the sphere @xmath137 ; over - dense and under - dense regions compensate each other when averaged over the entire surface of the sphere . the resulting potential fluctuations likewise average to zero over the sphere , as one can show by using gauss s theorem to evaluate the average gradient of the potential and integrating inward from @xmath138 . finally , a subtle bias arises if the particles used to generate the potential are also used to probe the potential , since local overdensities are sampled more heavily . to avoid this , the potential can be measured at a set of points @xmath139 which are _ independent _ of the particle positions @xmath7 . then @xmath140 should display some scatter , but average to zero when integrated over test points @xmath139 within a spherical shell . between @xmath0-body and smoothed potentials for tapered hernquist ( left ) and jaffe ( right ) models , normalized by central potential @xmath129 . vertical dashed lines show value of @xmath69 . points are results for uniform realizations ; jagged curves shows averages for groups of 32 points . grey - scale images show typical results for _ random _ realizations . central potentials are @xmath141 and @xmath142 . [ deltaphihj ] ] to 0.10 between @xmath0-body and smoothed potentials for tapered hernquist ( left ) and jaffe ( right ) models , normalized by central potential @xmath129 . vertical dashed lines show value of @xmath69 . points are results for uniform realizations ; jagged curves shows averages for groups of 32 points . grey - scale images show typical results for _ random _ realizations . central potentials are @xmath141 and @xmath142 . [ deltaphihj ] ] fig . [ deltaphihj ] shows results from direct - sum potential calculations for tapered hernquist and jaffe models , using units with @xmath143 . in each case , the underlying density profile was represented with @xmath144 equal - mass particles , and the potential was measured at @xmath145 points uniformly distributed in @xmath146 between @xmath147 and @xmath148 . the points show results for uniform radial sampling . while non - radial fluctuations create scatter in @xmath149 , the distribution is fairly symmetric about the line @xmath150 . the jagged curves are obtained by averaging @xmath151 over radial bins each containing @xmath152 points . these averages fall near zero , demonstrating very good agreement between the @xmath0-body results and the potentials calculated from the smoothed density profiles . for comparison , the grey - scale images in fig . [ deltaphihj ] display representative results for _ random _ realizations of each density profile . in these realizations , the radius of particle @xmath9 is computed by solving @xmath153 , where @xmath154 is a random number with @xmath155 yields systematically different @xmath134 values . the results shown here use the tausworthe generator ` taus2 ` @xcite . ] uniformly distributed between @xmath156 and @xmath68 . to examine the range of outcomes , @xmath157 random realizations of each model were generated and ranked by central potential @xmath134 ; since particle radii are chosen independently , the central limit theorem implies that @xmath134 has a normal distribution . shown here are the @xmath158 percentile and @xmath159 percentile members of these ensembles ; half of all random realizations lie between the two examples presented in each figure . note that these examples deviate from the true potential by fractional amounts of @xmath160 . obviously , it s impossible to detect discrepancies between @xmath149 and @xmath161 of less than one part in @xmath162 using random realizations with @xmath163 . it s instructive , not to mention disconcerting , to try reproducing fig . [ deltaphihj ] using a tree code @xcite instead of direct summation . tree codes employ approximations which become less accurate for @xmath164 @xcite ; these systematically bias computed potentials and accelerations ( see appendix b ) . for example , the code which will shortly be used for dynamical tests , run with an opening angle @xmath165 , yields central potentials which are too deep by a few parts in @xmath162 , depending on the system being modeled . this systematic error can not be ` swept under the carpet ' when comparing computed and predicted potentials at the level of precision attempted here . constructing equilibrium configurations is an important element of many @xmath0-body experiments . approximate equilibria may be generated by a variety of ad hoc methods , but the construction of a true equilibrium @xmath0-body model amounts to drawing particle positions and velocities from an equilibrium distribution function @xmath166 . however , a configuration based on a distribution function ( df ) derived without allowing for softening will _ not _ be in equilibrium if it is simulated with softening . assume the model to be constructed is spherical and isotropic . broadly speaking , there are two options : ( a ) adopt a df @xmath167 which depends on the energy @xmath168 , and solve poisson s equation for the gravitational potential , or ( b ) adopt a mass model @xmath169 , and use eddington s ( @xcite ) formula to solve for the df . if softening is taken into account , option ( a ) becomes somewhat awkward , since the source term for poisson s equation ( [ eq : smooth - poisson ] ) is non - local - body potentials which implements option ( a ) . on the other hand , option ( b ) is relatively straightforward ( e.g. , * ? ? ? starting with a desired density profile @xmath118 , the first step is to compute the smoothed density and mass profiles @xmath72 and @xmath120 , respectively . since @xmath170 everywhere , equation ( [ eq : potential - equation ] ) guarantees that the smoothed potential @xmath121 is a monotonically increasing function of @xmath20 . it is therefore possible to express the underlying density profile @xmath118 as a function of @xmath121 , and compute the df : @xmath171 note that in @xmath172 , the quantity @xmath169 is the _ underlying _ density , while @xmath173 is the _ smoothed _ potential , related by poisson s equation to the smoothed density @xmath72 . in effect , the smoothed potential @xmath121 is taken as a given , and ( [ eq : eddington - formula ] ) is used to find what will hereafter be called the _ smoothed distribution function _ @xmath174 ; with this df , the underlying profile @xmath118 is in equilibrium in the adopted potential ( e.g. , * ? ? ? * ; * ? ? ? conversely , setting @xmath12 yields the _ self - consistent distribution function _ @xmath175 which describes a self - gravitating model with the underlying profile @xmath118 . ; dashed , dot - dashed , and dotted curves : @xmath174 for @xmath66 , @xmath107 , and @xmath108 , respectively . [ dfhj ] ] to 0.10 ; dashed , dot - dashed , and dotted curves : @xmath174 for @xmath66 , @xmath107 , and @xmath108 , respectively . [ dfhj ] ] fig . [ dfhj ] presents dfs for tapered hernquist and jaffe models . in each case the solid line shows the self - consistent df ; these match the published dfs @xcite over almost the entire energy range , deviating only for @xmath176 where tapering sets in . smoothed dfs for jaffe models appear very different from their self - consistent counterpart . the latter has a logarithmic , infinitely deep potential well , which effectively confines material with constant velocity dispersion in a @xmath39 cusp . the characteristic phase - space density @xmath177 diverges as @xmath46 ( ie , as @xmath178 ) , but only because @xmath179 does . with @xmath164 the potential well is harmonic at small @xmath20 , and ca nt confine a @xmath39 cusp unless the local velocity dispersion scales as @xmath180 ; thus the phase - space density now diverges as @xmath181 . moreover , the domain of @xmath174 is limited to to @xmath182 , where @xmath183 . thus , instead of growing exponentially as a function of @xmath184 , the smoothed df abruptly diverges at some finite energy . by comparison , smoothed dfs for hernquist models look similar to the self - consistent df . the latter has a potential well of finite depth , and the smoothed profiles generate wells which are only slightly shallower . as the left panel of fig . [ dfhj ] shows , all the dfs asymptote to @xmath185 as @xmath186 . however , the run of velocity dispersion with @xmath20 is different ; the self - consistent model has @xmath187 , implying @xmath188 . in contrast , the smoothed models have @xmath180 , implying @xmath189 . one consequence is that the way in which @xmath190 as @xmath186 is different in the smoothed and self - consistent hernquist models . the self - consistent model has a linear potential as small @xmath20 , and thus @xmath191 . by comparison , the models based on smoothed potentials have harmonic cores , and as a result , @xmath192 . ( this difference is not apparent in fig . [ dfhj ] but becomes obvious when @xmath193 is plotted against @xmath194 . ) in this respect , the use of a smoothed potential effects a non - trivial change on hernquist models : @xmath195 is a different power - law of @xmath196 . coincidentally , the smoothed jaffe models have @xmath197 , just like the self - consistent hernquist model . @xmath0-body simulations are useful to show that the distribution functions just constructed are actually in dynamical equilibrium with their smoothed potentials . for each model and @xmath11 value , two ensembles of three random realizations were run . in one ensemble , the initial conditions were generated using the self - consistent df @xmath175 . the other ensemble used initial conditions generated from the smoothed df @xmath174 , which allows for the effects of softening . each realization contained @xmath198 equal - mass particles . initial particle radii @xmath128 were selected randomly by solving @xmath153 as described above . initial particle speeds @xmath199 were selected randomly by rejection sampling @xcite from the distributions @xmath200 or @xmath201 , where the former assumes the self - consistent df , and the latter a smoothed df . position and velocity vectors for particle @xmath9 are obtained by multiplying @xmath128 and @xmath199 by independent unit vectors drawn from an isotropic distribution . in effect , this procedure treats the 6-d distribution function @xmath202 as a probability density , and selects each particle s coordinates independently . simulations were run using a hierarchical @xmath0-body code . an opening angle of @xmath165 , together with quadrupole moments , provided forces with median errors @xmath203 . particles within @xmath204 have much larger force errors , although these seem to have limited effect in practice ( appendix b ) . trajectories were integrated using a time - centered leap - frog , with the same time - step @xmath205 for all particles ( see 4.3.1 ) . all simulations were run to @xmath206 , which is more than sufficient to test initial equilibrium . -body simulations of hernquist ( top row ) and jaffe ( bottom row ) models , each run with the @xmath11 value labeled . solid ( dotted ) curves show results for initial conditions generated from smoothed ( self - consistent ) dfs . light grey bands show expected @xmath207 variation in central potential for @xmath208 independent particles . [ phievol ] ] fig . [ phievol ] shows how the potential well depth @xmath209 of each simulation evolves as a function of time . here , well depth is estimated by computing the softened gravitational potential @xmath119 of each particle @xmath9 and taking the minimum ( most negative ) value . ( this is more accurate than evaluating the potential at the origin since the center of the system may wander slightly during a dynamical simulation . ) to better display the observed changes in @xmath209 , the horizontal axis is logarithmic in time . most of the ensembles set up without allowing for softening ( dotted curves in fig . [ phievol ] ) are clearly not in equilibrium . in all three of the jaffe models ( bottom row ) , the potential wells become dramatically shallower on a time - scale comparable to the dynamical time at @xmath210 . the reason for this is evident . the self - consistent jaffe model has a central potential diverging like @xmath211 as @xmath46 ; this potential can confine particles with finite velocity dispersion at arbitrarily small radii . however , the relatively shallow potential well of a smoothed jaffe model can not confine these particles ; they travel outward in a coherent surge and phase - mix at radii of a few @xmath11 . their outward surge and subsequent fallback accounts for the rapid rise and partial rebound of the central potential . similar although less pronounced evolution occurs in the hernquist models ( top row ) with @xmath212 and possibly with @xmath213 as well . only the self - consistent hernquist models run with @xmath214 appear truly close to equilibrium . in contrast , all of the ensembles set up with smoothed dfs ( solid curves in fig . [ phievol ] ) are close to dynamical equilibrium . in equilibrium , gravitational potentials fluctuate as individual particles move along their orbits . if particles are uncorrelated , the amplitude of these fluctuations should be comparable to the amplitude seen in an ensemble of independent realizations . to check this , @xmath157 realizations of each model were generated ; central gravitational potentials @xmath209 were evaluated using the same tree algorithm and parameters used for the self - consistent simulations . the grey horizontal bands show a range of @xmath207 around the average central potential for each model and choice of @xmath11 . some of the simulations set up using smoothed dfs wander slightly beyond the @xmath215 range . however , none of them exhibit the dramatic evolution seen in the cases set up using self - consistent dfs . for simulations run with @xmath66 , @xmath107 , and @xmath108 , respectively ; each is displaced downward by one additional unit in @xmath216 for clarity . [ rhoprofj ] ] the central potential is relatively insensitive to changes in the mass distribution on scales @xmath217 . to examine small - scale changes directly , density profiles measured from the initial conditions were compared to profiles measured at @xmath218 time units . these profiles were derived as follows . first , sph - style interpolation with an adaptive kernel containing @xmath152 particles was used to estimate the density around each particle . next , the centroid position @xmath219 of the @xmath152 highest - density particles was determined . finally , a set of nested spherical shells , centered on @xmath219 , were constructed ; each shell was required to contain at least @xmath220 particles and have an outer radius at least @xmath221 times its inner radius . [ rhoprofj ] summarizes results for jaffe models , which display the most obvious changes . the density of each shell is plotted against the average distance from @xmath219 of the particles it contains . in each panel , the top set of curves compare initial ( @xmath222 ) numerical results with the underlying tapered jaffe model , always represented by a light grey line . profiles from three independent @xmath0-body realizations of each model are overplotted . while some scatter from realization to realization is seen , the measured densities track the underlying profile throughout the entire range plotted . the outermost point is at @xmath223 times the radius of the innermost one ; there are not enough particles to obtain measurements at smaller or larger radii . ranged below the top curves in fig . [ rhoprofj ] are numerical results at @xmath206 for softening lengths @xmath66 , @xmath107 , and @xmath108 , each shifted downward by one more unit in @xmath216 . again , profiles from three independent simulations are overplotted to illustrate run - to - run variations . simulations set up using the self - consistent df ( left panel ) show significant density evolution ; their initial power - law profiles are rapidly replaced by cores of roughly constant density inside @xmath224 . in contrast , simulations set up using smoothed dfs ( right panel ) follow the initial profile down to @xmath225 ( although density evolution occurs on smaller scales ) . this shows that a careful set - up procedure can maintain the initial density profile even on scales much smaller than the softening length . a similar plot for the hernquist models confirms that most of these simulations start close to equilibrium . hernquist models set up using smoothed dfs do nt appear to evolve at all , although this statement should be qualified since the profiles of these models ca nt be measured reliably on scales much smaller than @xmath226 . models set up using the self - consistent df and run with @xmath212 undergo some density evolution ; their profiles fall below the underlying hernquist model at @xmath227 , although they continue rising to the innermost point measured . simulations run with @xmath69 or less display no obvious changes down to scales of @xmath226 . it appears that the jaffe models set up using smoothed dfs are not _ completely _ free of long - term evolution . the right - hand panel of fig . [ rhoprofj ] shows that the peak density as measured using a fixed number of particles falls by roughly an order of magnitude by @xmath206 . moreover , a close examination of fig . [ phievol ] turns some cases with a gradual decrease in potential well depth ; in the smoothed jaffe models with @xmath66 , for example , @xmath209 exhibits an upward trend of @xmath228 percent per unit time . this evolution may not be due to any flaw in the initial conditions ; the central cusps of such models , which are confined by _ very _ shallow harmonic potentials , are fragile and easily disrupted . there may be more than one mechanism at work here ; the rate of potential evolution appears to be inversely proportional to particle number @xmath0 , while the rate of density evolution is independent of @xmath0 . a full examination of this matter is beyond the scope of this paper . selecting the time - step @xmath229 for an @xmath0-body simulation is a non - trivial problem . while the choice can usually be justified post - hoc by testing for convergence in a series of experiments with different time - steps , it s clearly convenient to be able to estimate an appropriate @xmath229 a priori . a general rule governing such estimates is that the time - step should be smaller than the shortest dynamical time - scale present in the simulation . the central density @xmath230 of a smoothed density profile defines one such time - scale . within the nearly constant - density core of a smoothed profile , the local orbital period is @xmath231 ; this is the shortest orbital period anywhere in the system . numerical tests show the leap - frog integrator is well - behaved if @xmath232 ( conversely , it becomes unstable if @xmath233 ) . among the models simulated here , the jaffe model with @xmath214 has the highest smoothed central density ; for this model , @xmath234 . given this density , @xmath235 and @xmath236 . the time required for a fast - moving particle to cross the core region defines another time - scale . if @xmath237 is depth of the central potential well , the maximum speed of a bound particle is @xmath238 , and the core crossing time is @xmath239 . the smoothed jaffe model with @xmath66 has the deepest potential well . for this model , tests of the leap - frog with fast - moving particles on radial orbits show that @xmath240 yields good results , but time - steps a few times longer result in poor energy conservation as particles traverse the core region . ( the relationship between @xmath241 and the maximum acceptable time - step @xmath229 may be somewhat model - dependent . ) thus , for the jaffe model with @xmath66 , both the local criterion based on @xmath242 and the global criterion based on @xmath241 yield similar constraints , one can show that @xmath243 for any smoothed jaffe model ; both criteria yield similar constraints almost independent of @xmath11 . for smoothed hernquist models , on the other hand , @xmath244 ; the constraint based on @xmath241 is generally stricter . ] on @xmath229 . it s convenient to round @xmath229 down to the next power of two , implying @xmath205 . this corresponds to @xmath245 , which is somewhat conservative but helps insure that non - equilibrium changes will be accurately followed . to see if this time - step is reasonable , realizations of this model were simulated with various values of @xmath229 between @xmath67 to @xmath246 . at the lower end of this range , the effects of an over - large time - step manifest quickly ; global energy conservation is violated , and the measured central potential @xmath209 drifts upward over time ( even though the initial conditions , generated from @xmath174 , are near equilibrium ) . with a time - step @xmath247 , for example ; the potential well becomes @xmath248 percent shallower during the first two time - steps , and by @xmath249 its depth has decreased by @xmath250 percent . these simulations also violate global energy conservation , becoming @xmath251 percent less bound by @xmath249 . integration errors are reduced but not entirely eliminated with a time - step @xmath252 ; by @xmath249 , the potential well becomes @xmath253 percent shallower , while total energy changes by @xmath254 percent . with a time - step of @xmath255 or less , global energy conservation is essentially perfect , and variations in @xmath209 appear to be driven largely by particle discreteness as opposed to time - step effects . plots analogous to fig . [ rhoprofj ] show that the simulations with time - steps as large as @xmath256 reproduce the inner cusps of jaffe models just as well as those with @xmath205 . with this time - step , individual particles may not be followed accurately , but their aggregate distribution is not obviously incorrect . on the other hand , a time - step @xmath247 yields density profiles which fall below the initial curves for @xmath257 . softening and smoothing are mathematically equivalent . while the particular form of softening adopted in ( [ eq : nbody ] ) corresponds to smoothing with a plummer kernel ( [ eq : plummer - kernel ] ) , other softening prescriptions can also be described in terms of smoothing operations ( e.g. * ? ? ? there are two conceptual advantages to thinking about softening as a smoothing process transforming the underlying density field @xmath258 to the smoothed density field @xmath259 . first , since the gravitational potential @xmath260 is related to @xmath259 by poisson s equation , the powerful mathematical machinery of classical potential theory becomes available to analyze potentials in @xmath0-body simulations . second , focusing attention on smoothing makes the source term for the gravitational field explicit . from this perspective , smoothing is not a property of the particles , but a separate step introduced to ameliorate @xmath14 singularities in the potential . particle themselves are points rather than extended objects ; this insures that their trajectories are characteristics of ( [ eq : vlasov ] ) . plummer smoothing converts @xmath261 power - laws to cores . at radii @xmath262 the density profile is essentially unchanged , while at @xmath217 the density approaches a constant value equal to the density of the underlying model at @xmath263 times a factor which depends only on @xmath48 . for the case @xmath52 , this factor is unity and @xmath264 everywhere . the effects of plummer smoothing on astrophysically - motivated models with power - law cusps , such as the jaffe , hernquist , and nfw profiles , follow for the most part from the results for pure power - laws . in particular , for @xmath88 , where @xmath59 is the profile s scale radius , the power - law results are essentially ` grafted ' onto the underlying profile . on the other hand , for @xmath87 , the inner power - law is erased by smoothing . smoothing provides a way to predict the potentials obtained in @xmath0-body calculations to an accuracy limited only by @xmath126 fluctuations . these predictions offer new and powerful tests of @xmath0-body methodology , exposing subtle systematic effects which may be difficult to diagnose by other means . given an underlying density profile @xmath118 , it s straightforward to construct an isotropic distribution function @xmath174 such that @xmath118 is in equilibrium with the potential generated by its smoothed counterpart @xmath72 . such distribution functions can be used to generate high - quality equilibrium initial conditions for @xmath0-body simulations ; they should be particularly effective when realized with ` quiet start ' procedures @xcite . systems with shallow central cusps , such as hernquist and nfw models , may be set up fairly close to equilibrium without taking softening into account as long as @xmath11 is not too large . however , it appears impossible to set up a good @xmath0-body realization of a jaffe model without allowing for softening . it s true that realizations so constructed do nt reproduce the actual dynamics of the underlying models at small radii ( * ? ? ? * footnote 8) ; to obtain an equilibrium , the velocity dispersion is reduced on scales @xmath224 . but realizations set up _ without _ softening preserve neither the dispersion nor the density at small radii , and the initial relaxation of such a system ca nt be calculated a priori but must be simulated numerically . on the whole , it seems better to get the central density profile right on scales @xmath71 , and know how the central velocity dispersion profile has been modified . even if the dynamics are not believable within @xmath131 , the ability to localize mass on such scales may be advantageous in modeling dynamics on larger scales . mathematica code to tabulate smoothed models is available at + ` http://www.ifa.hawaii.edu/faculty/barnes/research/smoothing/ ` . i thank jun makino and lars hernquist for useful and encouraging comments , and an anonymous referee for a positive and constructive report . mathematica rocks . barnes , j.e . 1998 , ` dynamics of galaxy interactions ' , in _ galaxies : interactions and induced star formation _ , d. friedli , l. martinet , & d. pfenniger . berlin , springer , p. 275394 debattista , v.p . & sellwood , j.a . 2000 , ` constraints from dynamical friction on the dark matter content of barred galaxies ' , apj , * 543 * , 704721 the power - law density and cumulative mass profiles are @xmath265 plummer smoothing converts power laws with @xmath266 to finite - density cores . at @xmath217 the smoothed density is nearly constant and close to the smoothed central density @xmath267 . within this constant - density region , the smoothed mass profile is approximately @xmath268 at @xmath262 , on the other hand , smoothing has little effect on the mass profile , so @xmath269 . interpolating between these functions yields an approximate expression for the smoothed mass profile : @xmath270 where the shape parameter @xmath271 determines how abruptly the transition from one function to the other takes place . this expression can be rearranged to give @xmath272 the smoothed density profile is obtained by differentiating the mass profile : @xmath273 profile , computed for @xmath32 using ( [ eq : approx - smooth - mass ] ) and ( [ eq : approx - smooth - rho ] ) . dark curves show results for @xmath274 ; light grey solid curves show errors in density only for @xmath275 ( above ) and @xmath276 ( below ) . [ errorcusp2 ] ] to 0.10 profile , computed for @xmath32 using ( [ eq : approx - smooth - mass ] ) and ( [ eq : approx - smooth - rho ] ) . dark curves show results for @xmath274 ; light grey solid curves show errors in density only for @xmath275 ( above ) and @xmath276 ( below ) . [ errorcusp2 ] ] figs . [ errorcusp1 ] and [ errorcusp2 ] present tests of these approximations for @xmath31 and @xmath277 power - laws , respectively . as in figs . [ cusp1 ] and [ cusp2 ] , the smoothed density profile was computed with @xmath32 ; for other values of @xmath11 , the entire pattern simply shifts left or right without changing shape or amplitude . dashed curves show relative errors in smoothed mass , @xmath278 , while solid curves are relative errors in smoothed density @xmath279 . the @xmath271 value used for each dark curve is the value which minimizes @xmath280 evaluated at points @xmath128 distributed uniformly in @xmath211 between @xmath281 and @xmath282 . in light grey , plots of @xmath283 for two other @xmath271 illustrate the sensitivity to this parameter . comparing these plots , it appears that the approximation works better for the @xmath31 power - law than it does for @xmath277 , but even in the latter case the maximum error is only @xmath284% . because ( [ eq : approx - smooth - mass ] ) modifies the underlying mass profile with a multiplicative factor , it can also be used to approximate effects of softening on non - power - law profiles ( e.g. * ? ? ? * ) ; for this purpose , both @xmath11 and @xmath271 can be treated as free parameters and adjusted to provide a good fit . the resulting errors in density , which amount to a few percent near the softening scale , are undesirable but do nt appear to seriously compromise @xmath0-body simulations with @xmath285 . for @xmath84 , smoothing primarily modifies the @xmath286 part of these density profiles . this , together with the exact solution for the case @xmath19 given in 2.1 , suggests simple approximations for smoothed hernquist and nfw models : @xmath287 fig . [ errorrhohnfw ] plots the relative error in density , @xmath288 for both models , adopting @xmath289 . for other values of @xmath11 , these errors scale roughly as @xmath290 . the general behavior of these approximations is readily understood . overall , @xmath291 is more accurate than @xmath292 since the nfw profile is closer to @xmath31 at all radii . both curves are approximately flat for @xmath217 , then reach minima for @xmath20 between @xmath11 and the profile scale radius @xmath59 . these minima arise because the smoothed density approaches or even slightly exceeds the underlying density ( see figs . [ rhoh ] and [ rhonfw ] ) , while ( [ eq : approx - rho - cusp1 ] ) always yields values below the underlying density . it s sometimes useful to have the cumulative mass for a smoothed profile . the approximate profiles in ( [ eq : approx - rho - cusp1 ] ) can be integrated analytically , although the resulting expressions are a bit awkward : @xmath293 and @xmath294 note that because the approximate profiles ( [ eq : approx - rho - cusp1 ] ) systematically underestimate the true smoothed densities , these expressions will likewise systematically underestimate the total smoothed mass . , plotted vs. radius @xmath20 , for the approximations given in ( [ eq : approx - rho - cusp1 ] ) . solid and dashed curves show results for hernquist and nfw profiles , respectively , computed for @xmath289 . [ errorrhohnfw ] ] tree codes reduce the computational cost of gravitational force calculation by making explicit approximations @xcite . the long - range potential due to a localized mass distribution @xmath295 with total mass @xmath60 and center of mass position @xmath296 is approximated as @xmath297 where the higher order terms include quadrupole and possibly higher - order moments ( dipole terms vanish because @xmath296 coincides with the center of mass ) . to implement softening , this approximation is typically replaced with @xmath298 this works at large distances , but becomes inaccurate if @xmath299 @xcite . moreover , because the error is introduced at the _ monopole _ level @xcite , higher - order corrections do nt repair the damage . to appreciate the problem , consider a sphere @xmath300 centered on @xmath296 with radius @xmath42 large enough to enclose @xmath295 . for @xmath12 , the inward acceleration averaged over the surface of @xmath300 is easily computed using gauss s theorem : @xmath301 in other words , the monopole term is sufficient to calculate the inward acceleration averaged over the surface of @xmath300 _ exactly_. suppose we want to compute @xmath302 for @xmath164 . again using gauss s theorem , we have @xmath303 is the smoothed mass within the sphere . as before , this is an exact equality . the tree code approximation ( [ eq : approx - pot - soft ] ) implies that the enclosed mass is @xmath304 this is correct if @xmath295 is simply a point mass located at @xmath296 . but if @xmath295 has finite extent , then the enclosed mass @xmath305 is _ always _ less than @xmath306 . as a result , ( [ eq : approx - pot - soft ] ) will systematically _ overestimate _ the inward acceleration and depth of the potential well due to @xmath295 . @xcite demonstrate a similar result by computing the softened potential of a homogeneous sphere ; they find @xmath307 is only the first term in a series . the inequality @xmath308 is easily verified for plummer softening . an analogous inequality is likely to hold for other smoothing kernels @xmath309 which monotonically decrease with @xmath20 . smoothing kernels with compact support @xcite may be better behaved in this regard . under what conditions are these errors significant ? for ` reasonable ' values of @xmath11 , most dynamically relevant interactions are on ranges @xmath310 where softening has little effect ; these interactions are not compromised since ( [ eq : approx - pot - soft ] ) is nearly correct at long range . only if a significant fraction of a system s mass lies within a region of size @xmath11 can these errors become important . this situation was not investigated in early tree code tests ( e.g. , * ? ? ? * ; * ? ? ? * ) , which generally used mass models with cores instead of central cusps , and even heavily softened hernquist models do nt have much mass within one softening radius . on the other hand , jaffe models pack more mass into small radii ; a jaffe model with @xmath289 has almost @xmath311 percent of its mass within @xmath210 . jaffe models should be good test configurations for examining treecode softening errors . tests were run using tapered jaffe and hernquist models , realized using the same parameters ( @xmath109 , @xmath110 , @xmath111 , and @xmath312 ) used in the dynamical experiments ( 4.3 ) . in each model , the gravitational field was sampled at @xmath145 points drawn from the same distribution as the mass . at each test point , results from the tree code with opening angle @xmath165 , including quadrupole terms , were compared with the results of an direct - sum code . as expected , the tests with hernquist models showed relatively little trend of force calculation error with @xmath11 , although the errors are somewhat larger for @xmath212 than for smaller values . in contrast , the tests with jaffe models reveal a clear relationship between softening length and force calculation accuracy . for simulations run with @xmath165 and @xmath313 . light grey , dark grey , and black show potential differences for @xmath66 , @xmath107 , and @xmath108 , respectively ; three independent realizations are plotted in each case . [ phidiffj ] ] for simulations run with @xmath165 and @xmath313 . light grey , dark grey , and black show potential differences for @xmath66 , @xmath107 , and @xmath108 , respectively ; three independent realizations are plotted in each case . [ phidiffj ] ] @xmath11 . here @xmath315 and @xmath316 are accelerations computed using a tree code and direct summation , respectively . the grey dots represent measurements of @xmath317 for individual test points , computed using @xmath69 . the pattern of errors suggests two regimes . at radii @xmath318 ( @xmath319 ) , the points fall in a ` sawtooth ' pattern which reflects the hierarchical cell structure used in the force calculation . at smaller radii , on the other hand , the relative error grows more or less monotonically as @xmath46 . it appears that errors in the large-@xmath20 regime are due to neglect of moments beyond quadrupole order in computing the potentials of individual cells ; conversely , the errors in the small-@xmath20 regime are due to the tree code s inaccurate treatment of softening . the direction of the error vectors @xmath320 supports this interpretation ; in the large-@xmath20 regime they are isotropically distributed , while in the small-@xmath20 regime they point toward the center of the system . the jagged line threading through the dots in fig . [ accelerrorj ] , constructed by averaging test points in groups of @xmath220 , shows the overall relationship between acceleration error and radius for @xmath69 . similar curves are also plotted for @xmath212 ( above ) and @xmath66 ( below ) . at large radii , all three curves coincide precisely , implying that force errors are independent of @xmath11 . going to smaller @xmath20 , the curve for @xmath212 is the first to diverge , rising above the other two , next the curve for @xmath213 begins to rise , tracking the mean distribution of the plotted dots , and finally the curve for @xmath66 parallels the other two . each curve begins rising monotonically at a radius @xmath321 ; this is evidently where softening errors begin to dominate other errors in the force calculation . at the softening radius @xmath210 , all three curves show mean acceleration errors @xmath322 . are errors of this magnitude dynamically important ? in particular , could they explain some of the potential evolution seen in the runs set up with softening ( solid curves ) in fig . [ phievol ] ? one possible test is to re - run the simulations using smaller @xmath323 values ; decreasing @xmath323 from @xmath324 to @xmath325 or @xmath313 reduces the tree code acceleration error associated with softening by factors of @xmath284 or @xmath248 , respectively ( albeit at significant computational costs ) . the new runs were started using exactly the same initial conditions as their @xmath165 counterparts . this initially allows central potentials @xmath326 to be compared between runs to high accuracy , temporarily circumventing the effects of @xmath126 fluctuations . [ phidiffj ] compares jaffe model results for @xmath165 and @xmath313 . initially , the potential difference @xmath327 arises because the treecode systematically overestimates the potential well depth by an amount which is greater for larger values of @xmath323 . as the simulations run , the excess radial acceleration causes systems run with @xmath165 to contract relative to those run with @xmath328 , causing a further increase in potential well depth . this contraction takes place on a dynamical time - scale at @xmath329 , occurring first for the @xmath66 simulations ( light grey curves ) . at later times , as trajectories in otherwise - identical simulations with different @xmath323 values diverge , short - term fluctuations in @xmath209 de - correlate and the dispersion in potential differences increases markedly . these results show that force - calculation errors can have measurable , albeit modest , effects on dynamical evolution . extrapolating results for a range of @xmath323 values to @xmath330 , it appears that for @xmath165 the _ dynamical _ response of these jaffe models due to excess radial acceleration deepens central potential wells by about @xmath331 percent . to test this extrapolation , it would be instructive to repeat these experiments with a direct - summation code . however , compared to the overall range of @xmath209 variations seen in fig . [ phievol ] , the perturbations due to force calculation errors seem relatively insignificant . in addition , density profiles measured from runs with @xmath332 or @xmath313 appear similar to those shown in fig . [ rhoprofj ] , again implying that treecode force calculation errors have little effect on the key results of this study . the best way to correct ( [ eq : approx - pot - soft ] ) for short - range interactions is not obvious . a simple , ad - hoc option is to reduce the effective opening angle on small scales . for example , accepting cells which satisfy @xmath333 where @xmath334 is the distance to the cell s center of mass , @xmath335 is the cell s size , measured along any edge , and @xmath336 is the distance between the cell s center of mass and its geometric center @xcite , reduces softening - relating treecode errors by a factor of @xmath284 at a modest cost in computing time . further experiments with similar expressions may produce better compromises between speed and accuracy .
in self - consistent @xmath0-body simulations of collisionless systems , gravitational interactions are modified on small scales to remove singularities and simplify the task of numerically integrating the equations of motion . this ` gravitational softening ' is sometimes presented as an ad - hoc departure from newtonian gravity . however , softening can also be described as a smoothing operation applied to the mass distribution ; the gravitational potential and the smoothed density obey poisson s equation precisely . while ` softening ' and ` smoothing ' are mathematically equivalent descriptions , the latter has some advantages . for example , the smoothing description suggests a way to set up @xmath0-body initial conditions in almost perfect dynamical equilibrium . methods : numerical galaxies : kinematics & dynamics
despite the spectacular success of quantum mechanics as a fundamental theory and as a superb calculational tool for all modern physics , the scientific literature continuously carries a substantial number of papers devoted to the foundations and the interpretation of the theory . the introduction of probability in the sense used in quantum mechanics has been the key issue of debates . since its early days the probabilistic interpretation of quantum phenomena has had its opponents , and einstein in a letter to born expressed the view @xcite : 0.3 cm _ the theory produces a good deal but hardly brings us closer to the secret of the old one . i am at all events convinced that _ he _ does not play dice . _ 0.3 cm this statement is the best known example of the open criticism of the probabilistic interpretation of quantum mechanics . in his fundamental paper devoted to the quantum theory of spontaneous and stimulated processes by radiation , einstein expressed his well - known discomfort about the probabilistic description of spontaneous emission in the following words @xcite : 0.3 cm _ the weakness of the theory lies on the one hand in the fact that it does not get us any closer to making the connection with wave theory ; on the other hand , that it leaves the duration and direction of the elementary process to chance. _ 0.3 cm in making the connection between waves and particles , einstein noted that the emission of radiation ( outgoing radiation ) is directional . in the same paper he wrote : 0.3 cm _ outgoing radiation in the form of spherical waves does not exist . _ 0.3 cm in the early 1920 s , einstein , in his unpublished speculations , proposed the idea of a `` gespensterfeld '' or a ghost field which determines the probability for a light - quantum to take a definite path . in these speculations , the ghost field gives the relation between a wave field and a light - quantum by triggering the elementary process of spontaneous emission . the directionality of the elementary process is fully described by the will ( dynamical properties ) of the ghost field ( see fig . [ fig1 ] ) . the probabilistic interpretation of quantum mechanics as developed by born has been considerably influenced by the idea of the ghost field @xcite . born realized that unlike the electromagnetic field , the schrdinger wave function @xmath1 has no direct physical reality . only if one interprets @xmath2 or the ghost field as a real probability , do all the weaknesses of the interpretation disappear because : @xmath3 in a paper presented at oxford , born expressed the opinion that @xcite : 0.3 cm _ but , of course , anybody dissatisfied with these ideas may feel free to assume that there are additional parameters not yet introduced into the theory which determine the individual event . _ 0.3 cm these additional parameters called , accordingly , hidden variables and denoted by @xmath4 , have been introduced in order to remove from quantum mechanics the indeterminacy introduced by the born interpretation of the wave function . according to hidden variables theories , all physical observables such as , for example the position and the momentum of a particle , the spin or the electric field are entirely deterministic , i.e , are objective realities given by functions @xmath5 . due to the assumed unobservable nature of the hidden parameters @xmath4 , only the following average @xmath6 can be measured in realistic experimental situations . the distribution of the as yet undiscovered hidden variables is positive ( a well - behaved probability distribution ) and normalized @xmath7 in this interpretation the hidden parameters would be equivalent to a ghost field of some yet undisclosed nature , i.e. , @xmath8 in 1935 , einstein , podolsky , and rosen ( epr ) published their famous paper @xcite in which they argued that : 0.3 cm _ the wave function @xmath1 does not provide a complete description of physical reality . _ 0.3 cm the need for a revision of the interpretation of the wave function was strongly expressed in this paper . in fact , epr questioned the completeness of quantum mechanics , analyzing a correlated system of two particles . the epr arguments can be best summarized in bohm s version @xcite which uses two correlated spin-@xmath0 particles as the epr physical system . let us consider an initial physical system consisting of two distinguishable spin-@xmath0 particles @xmath9 and @xmath10 in a spin-@xmath11 quantum state ( a spin singlet state ) . this system dissociates to a pair of spatially separated spin-@xmath0 particles . let assume that the dissociation process conserves the total spin of the system and that the separated particles are now moving to opposite stern - gerlach detectors ( fig . [ fig2 ] ) . in the framework of quantum mechanics the bohm singlet wave function corresponding to the epr dissociating system of two spin-@xmath12 particles @xmath9 and @xmath10 is : @xmath13 are the two eigenstates of the pauli matrix @xmath14 ( quantum operators are denoted with carets ) corresponding to the two eigenvalues @xmath15 and @xmath16 . after the dissociation the particles @xmath9 and @xmath10 move to opposite directions without ever being disturbed in any way and various spin components of each of these particles may by measured independently . the particle @xmath9 is detected by a stern - gerlach apparatus which can be oriented along directions @xmath17 , @xmath18 , and @xmath19 respectively . the particle @xmath10 is detected by a stern - gerlach apparatus which can be oriented along the same three directions @xmath17 , @xmath18 , and @xmath20 respectively ( fig . [ fig2 ] ) . the epr argument is based on the following observation involving such measurements . the stern - gerlach apparatus may record that the spin @xmath9 is parallel or antiparallel to @xmath17 , which one can take to be the z - axis . let us assume that it has been detected as parallel to @xmath17 meaning that we know that the value of the spin @xmath9 is @xmath21 . knowing that the spin of particle @xmath9 is @xmath15 , common sense based on an objective reality associated with the spin ( a singlet state consists of two antiparallel spins ) predicts that the second particle @xmath10 has its spin @xmath16 . this means that that the result of a measurement carried on the second particle by a stern - gerlach apparatus with the same orientation @xmath17 will predict an antiparallel spin . the same argument based on objective realities for spin variables can be repeated for the orientations @xmath18 and @xmath19 of the stern - gerlach apparatus leading to the conclusion that the values of the spin are known in the three directions which is in violation of quantum mechanical properties of pauli spin operators projected along @xmath17 , @xmath18 , and @xmath19 respectively . accordingly epr concluded that one should search for a deeper - lying theory in which the following two very important criteria are satisfied : * * the reality criterion * : _ if without in any way disturbing the system , we can predict with certainty the value of a physical quantity , then there exists an element of physical reality corresponding to this physical quantity . _ * * the locality criterion * : _ if at the time of measurement ... two systems no longer interact , no real change can take place in the second system in consequence of anything that may be done to the first system . _ the epr analysis of quantum correlations based on the reality and the locality criteria have led to apparently contradictory conclusions with quantum mechanics which do not allow for commuting objective realities associated with the spin observables . epr summarized their views in the following statement : 0.3 cm _ this simultaneous _ [ predictability ] _ makes the reality of _ [ the spins in bohm s version ] _ dependent upon the process of measurement carried out on the first system which does not disturb the second system in any way . no reasonable definition of reality could be expected to permit this . _ 0.3 cm if the spin singlet wave function ( [ singlet ] ) of the two particles is measured by two stern - gerlach analyzers , in the directions given by unit vectors @xmath17 and @xmath18 , the single detection of the spins is denoted by @xmath22 and @xmath23 , and the joint simultaneous detection of the particles @xmath9 and @xmath10 is denoted by @xmath24 . these expectation values for this wave function ( [ singlet ] ) are given by the following formulas : @xmath25 where @xmath26 and @xmath27 are the projections of the three spin-@xmath12 pauli matrices of the particle @xmath9 and @xmath10 on the stern - gerlach apparatus orientations characterized by the unit vectors @xmath17 and @xmath18 . quantum mechanics predicts that for parallel detectors we have a perfect correlation : @xmath28 is @xmath29 and as a result we have with probability one that the eigenvalue of @xmath30 is @xmath31 . this perfect correlation between these two measurements permits one to predict the outcome of the second measurement with probability * one * without ever disturbing the second system . on other hand in the framework of hidden variables or an unknown ghost field these expectation values are given by the following expression @xcite : @xmath32 in this formula @xmath33 describes a joint probability distribution of the ghost field characterized by the hidden variables @xmath34 and @xmath35 . the objective realities of the spin variables are given by the deterministic functions @xmath36 and @xmath37 . the locality assumption leads to the property that the distribution function of the ghost field is independent of the surrounding experimental setup and does not depend on what other observables are measured simultaneously on the same system . it is the purpose of this article to provide a quantum mechanical description of the ghost field . using different descriptions and formulations of quantum mechanics , we shall discuss the same epr correlation in different , though , equivalent forms of quantum theory . in short , we shall allow quantum mechanics to speak for itself on such issues as locality and the ghost field . we believe that letting quantum mechanics speak out via its different formulations can provide the most down - to - earth description of these fundamental and puzzling issues . as a result , our approach will be almost completely divorced from all of its philosophical or speculative aspects . instead of this , we shall focus our effort on the question : how can we make quantum mechanics look like a ghost field and vice versa ? these considerations demonstrate a much closer correspondence between the probabilistic interpretation of quantum mechanics and hidden variables or ghost fields . in our approach we shall show that the wave function @xmath1 can always be replaced by a suitable quantum mechanical ghost field @xmath38 with properties that defy the epr concepts of objective realities or locality . @xmath39 expressing the formalism of quantum mechanics in a given and suitable form can provide a better understanding of the nature of the quantum mechanical ghost field . instead of being vague in most of such descriptions , the formalism of quantum mechanics can provide a reasonable ( if not ideal ) description of quantum locality , reality , nonlocality and their interrelations . the concept of local realism as introduced by epr is based on the fundamental assumption that physical systems can be described by local objective properties that are independent of observation . starting with the epr article , questions devoted to local realism , to measurements and the nonlocal character of quantum correlations , and to hidden variables have been raised , debated and analyzed both theoretically and experimentally . this debate has produced various alternative theories to quantum mechanics . these theories have been considered over the years , hidden variables theories being prominent among these . from the theoretical point of view , a real breakthrough came after bell discovered@xcite in 1965 that certain classes of hidden variables theories limit the strength of spin or photon correlations and that quantum mechanical correlations are not constrained by these limits . these fundamental limitations or constrains inhibited by local realism lead to the famous bell inequality . for the purpose of proving the bell inequality we shall assume that stern - gerlach apparatuses for the two particles are oriented along the three directions @xmath17 , @xmath18 and @xmath19 with angles @xmath40 ( see fig . [ fig2 ] ) . in order to prove the bell inequality we introduce ( following cover @xcite ) the following quantity built from objective spin realities : @xmath41 we shall restrict the range of these objective realities to @xmath42 and @xmath43 of the stern - gerlach apparatus . because the spins are antiparallel we have with certainty that : @xmath44 for all the three orientations and for all the possible hidden variables @xmath34 and @xmath35 forming the ghost field that governs the statistical properties of the local objective realities representing the spin . this follows from the fact that if the first stern - gerlach apparatus oriented at @xmath17 , @xmath18 and @xmath20 detects the spin direction to be @xmath29 , the second stern - gerlach apparatus oriented as in fig . [ fig2 ] has to detect the other spin of the singlet state ( [ singlet ] ) as @xmath31 . from the fact that the objective realities are bounded by @xmath29 we note that ( [ sigma ] ) reaches a minimum value : @xmath45 from the fact that the spins are antiparallel i.e. , ( [ perfect ] ) holds , we have the following inequality involving the objective realities : @xmath46 multiplying this inequality by @xmath47 , and integrating over the ghost field variables we obtain the following bell inequality : @xmath48 where we have used the fact that correlations are symmetric i.e. , for example @xmath49 . in order to disprove a theory based on local objective realities a _ series _ of three of different correlation experiments is required ( in our case correlations @xmath17 @xmath18 , @xmath17 @xmath19 and @xmath18 @xmath19 ) . what we have seen in the process of deriving the bell inequality ( [ bell ] ) is that one could provide a much more accurate description of locality , local reality and correlations in terms of ( [ lhv ] ) . what is even more important in this inequality is the fact that all these ideas could be put to experimental verifications . for the three orientations from fig . [ fig2 ] the inequality ( [ bell ] ) reduces to @xmath50 and for the quantum correlation ( [ qepr ] ) we have : @xmath51 which is false and clearly violates the bell inequality ( [ bell ] ) for the spin-@xmath52 singlet state given by ( [ singlet ] ) . these inequalities have helped to put a large class of hidden variables theories to experimental tests @xcite . theories based on local realism have all failed these tests . so far , all the experiments involving correlations constrained by the bell s inequality have been positive for quantum mechanics and negative for various versions of local - realism @xcite . if only one correlation experiment is performed the odds against objective realities are not very good . even if disproved in a series of three experiments , objective realities leave a puzzling question . if they come so close to the quantum predictions , perhaps there are some elements of truth in their descriptions given by ( [ lhv ] ) . if so , which particular assumptions of a local hidden variable theory are in agreement or disagreement with quantum mechanics . in order to answer such a question , we shall describe in the following sections , genuine quantum correlations in a form which bears a lot of similarities to the objective reality description . once quantum mechanics has been formulated in the form expressed by eq . ( [ lhv ] ) a genuine effort can be made to address such issues as hidden variable objective realities and locality . once this relation is established , the question why quantum mechanics violates the bell inequality can be posed and investigated . the classical malus law predicts an attenuation of an incident polarized light beam with intensity @xmath53 , passing through a linear polarizer to obey @xmath54 . this attenuation depends on the relative angle @xmath55 between the polarization direction @xmath56 of the incoming wave and the orientation @xmath17 of the polarizer , i.e , @xmath57 . if the incoming light beam consists of a statistical mixture of polarized light , the incoming intensity is : @xmath58 in this formula , the integration is over all possible angles of the random polarization direction @xmath56 and the classical distribution function @xmath59 characterizes the statistical properties of the incident light beam polarization . in quantum mechanics a similar malus law holds for spin-@xmath0 particles detected by a stern - gerlach apparatus . let assume that the spin of the particle @xmath9 is oriented along an arbitrary unit direction @xmath60 . the quantum expectation value of such a spin , if detected by the stern - gerlach apparatus oriented in the direction @xmath17 is : @xmath61 in this expression @xmath62 is the relative angle between the orientation of the detected spin state @xmath63 and the stern - gerlach apparatus @xmath64 . the corresponding expectation value of the spin-@xmath0 projection operator @xmath65 is : @xmath66 is just the quantum version of the malus law for spin . light ( [ malus ] ) and spin-@xmath0 ( [ malus2 ] ) differ in this formulation by a factor @xmath0 in the relative angle involved in the malus law . following the classical malus law for an unpolarized light beam ( [ malus ] ) , one can write the following quantum spin correlation function for an arbitrary quantum state of the two spin-@xmath0 particles detected by the stern - gerlach apparatus : @xmath67 where the distribution function @xmath68 describes the quantum mechanical ghost field for an arbitrary state @xmath69 of the two spin-@xmath0 particles . this quantum mechanical expectation value ( [ qmalus ] ) has a remarkable similarity to the hidden variable equation ( [ lhv ] ) if the following correspondence is made @xcite : @xmath70 i.e. , the @xmath71 and @xmath72 are now the directional cosine functions between the directions of the polarizers @xmath17 and @xmath18 and the `` hidden - variable '' directions @xmath73 and @xmath74 averaged with respect to a local , i.e. , stern - gerlach apparatus independent ghost field @xmath75 . it is then clear that the quantum mechanical formula ( [ qmalus ] ) has the form of a hidden variable theory with the local spin realities given by @xmath76 and @xmath77 , and with a local ghost field represented by the distribution @xmath78 . for the epr correlated system of the two spin-@xmath0 systems , it is natural just to take the probability distribution in the form : @xmath79 where the two - dimensional dirac delta function indicates that the direction @xmath60 is antiparallel to the direction @xmath80 ie . , @xmath81 . the marginal properties of this distribution represent just an isotropic uniform distribution of the spin directions on the sphere : @xmath82 from these relations we conclude that individual ghost fields of the spins correspond to a uniform distribution of orientations . this leads to single spin averages equal to zero , i.e. , : @xmath83 on the other hand a direct integration of the malus correlation ( [ qmalus ] ) with the antiparallel distribution function given by eq . ( [ lgh ] ) leads to : @xmath84 this result does not violate the bell inequality ( [ bell ] ) and is off from the quantum mechanical result ( [ qepr ] ) by a factor of @xmath85 . this can be simply fixed by a change of the @xmath86-prefactor in ( [ lgh ] ) into @xmath87 . but this change has tremendous consequences for the normalization of the distribution function ( [ lgh ] ) . in order to preserve the normalization we write : @xmath88 we check that indeed with this change : @xmath89 we calculate the marginals : @xmath90 clearly the function ( [ qlgf ] ) has the correct marginals ( [ marginals ] ) , reproduces the right quantum mechanical prefactor and yet is not positive ! the appearance of such a nonpositive probability distribution function ( called for this reason a quasi - distribution function ) is a typical quantum property , whenever a probabilistic phase - space is constructed for noncommuting quantum observables . the first nonpositive quasi - probability distribution was introduced into quantum mechanics by wigner in 1932 ( @xcite ) for the position and the momentum phase - space . the quasi - distribution function ( [ qlgf ] ) corresponds to a wigner - type distribution function for the spin-@xmath0 phase - space variables . in fact it is possible to study and derive rigorously @xcite all kinds of quantum mechanical spin-@xmath0 quasi - distribution functions with ( [ qlgf ] ) being just an example corresponding to the quantum ghost field of the epr singlet state ( [ singlet ] ) . the quantum malus law ( [ qmalus ] ) , if applied to the joint correlations involving two stern - gerlach detectors , has the formal structure of a hidden variable theory , with the hidden parameters represented by `` hidden directions '' @xmath60 and @xmath91 . there the analogy ends because the quantum distribution corresponding to the local ghost field is a non - positive function that leads to the failure of the bell s inequalities for such a correlated state ( see fig . [ fig3 ] ) . using the malus form of spin correlations we have : @xmath92 in this section we are going to show that epr correlations can be described by a quantum nonlocal ghost field similar to the one used in a hidden - variables theory . we can rewrite the spin correlation function ( [ qepr ] ) in terms of a ghost field if we use the following two integral identities for the pauli matrices matrices : @xmath93 introducing these identities into ( [ qepr])we obtain : @xmath94 where the distribution function is given by the following quantum mechanical average : @xmath95 with the help of this distribution function we have rewritten the quantum mechanical correlation function in a form which has remarkable similarities to the correlation function given by eq . ( [ lhv ] ) . because the spin operators @xmath96 and @xmath97 have eigenvalues equal to @xmath15 or @xmath16 , i.e. , can represent only `` parallel '' ( `` p '' ) or `` antiparallel '' ( `` a '' ) outcomes , @xmath98 and @xmath99 can take only values equal to @xmath15 and @xmath16 . the bivalued distribution given by eq . ( [ ccd ] ) is positive everywhere and normalized : @xmath100 this function depends on the polarization directions @xmath17 and @xmath18 . the distribution function which depends on the orientation @xmath17 of the first stern - gerlach apparatus and on the orientation @xmath18 of the second ( possibly even remote ) stern - gerlach apparatus is nonlocal . because of this property we shall call an analyzer - dependent distribution function a nonlocal distribution function . the nonlocality of this distribution function makes the bell s inequality void , because in order to obtain this inequality the existence of a universal , local ( polarization independent ) distribution in the parameters @xmath98 and @xmath99 ( hidden parameters in this case ) like that appearing in ( [ lhv ] ) was essential @xcite . quantum mechanics tells us that if we insist on a distribution of the form given by eq . ( [ ccd ] ) , we can do it but only under the condition that the statistical distribution of the ghost field with the parameters @xmath98 and @xmath99 is nonlocal @xcite . this joint distribution function ( [ ccd ] ) can be written in the following form ( bayes theorem ) : @xmath101 where the distribution @xmath102 is the conditional of the event @xmath98 to occur under the condition that @xmath99 has occurred . the distribution @xmath103 is a one - fold marginal of ( [ ccd ] ) and is : @xmath104 this means the the one - fold marginal is local ( apparatus independent ) and that the outcomes are equally probable ie . , @xmath105 and @xmath106 . in order to elucidate the nonlocal distribution further we shall rewrite the nonlocal conditional distribution function in the following matrix form : @xmath107 from eq.([ccd ] ) and the form of the epr singlet state we obtain that the conditional probabilities for the particular transitions are @xmath108 this result shows that one can regard the epr correlations as just correlations of two sequences of random numbers jumping between values @xmath15 and @xmath16 ( `` p '' and `` a '' answers ) for polarization measurements performed with linear analyzers . the distribution @xmath109 is the conditional of the event @xmath34 ( `` p '' or `` a '' ) to occur under the condition that @xmath35 ( `` a '' or `` p '' ) has occurred . this positive and nonlocal distribution leads to a simple statistical interpretation of the spin transitions and of the violation of bell s inequality in terms of random numbers @xmath29 for the variables @xmath34 and @xmath35 . the quantum mechanical average in this case is represented by an ensemble average of two sequences of random numbers `` a '' and `` p '' . the random character of these variables can be applied to the description of the epr correlations measured by two polarizers . to each polarizer there corresponds a sequence of random variables denoted by @xmath34 and @xmath35 . these are the only possible outcome of the experiment . on each single polarizer the outcomes are completely random and the `` p '' and `` a '' answers occur with equal probability ( @xmath12 in this case ) . the nonlocality of the epr correlations shows up in the fact that these two perfectly random sequences ( on the first and the second analyzer ) are correlated and the correlations are given by eq . ( [ qepr ] ) . these formulas predict that the epr wave function can be understood as a nonlocal correlation between two random sequences @xmath110 and @xmath111 . the nonlocality of these correlations follows from the fact that whenever @xmath112 or @xmath113 on the polarizer @xmath18 , we must have @xmath114 or @xmath115 on the analyzer @xmath17 with the probability @xmath116 , i.e. , the outcomes on @xmath17 ( possibly even a remote analyzer ) are determined by the outcomes on the analyzer @xmath117 . for antiparallel correlations @xmath118 and the conditional matrix has the following form : @xmath119 this ghost field correspond to the following two outcomes at the stern - gerlach analyzers : @xmath120 we see that the outcomes correspond to two random sequences with perfect correlation corresponding to @xmath121 with conditional probability one and @xmath122 and @xmath123 with conditional probability zero . let us illustrate , using these random sequences , the violation of bell s inequality . let us assume that in the first series of experiments we set @xmath124 and @xmath125 . according to the formulas ( [ qnlgf ] ) we have for these angles : @xmath126 this means that with a 75% confidence we shall have that the outcomes on @xmath17 will be the same as the outcomes on @xmath18 and a 25% confidence that the outcomes are different . this leads to the spin correlation : @xmath127 we already have shown ( [ violation ] ) that these correlation do violate the bell inequality ( [ bell ] ) . the nonlocal ghost field contradicts the idea of objective realities because the probability of reproducing a sequence on the polarizer @xmath17 if the sequence on the polarizer @xmath18 is known is nonlocal if the ghost field distribution depends on the relative orientation @xmath55 of the polarizers . this is how the epr quantum nonlocal probabilities violate local realism ( fig . [ fig4 ] ) : @xmath128 in fact it is easy to give an example of a ghost field that will not violate the bell inequality in this case . if we select : @xmath129 we obtain that : @xmath130 which does not violate the bell inequality . the quantum nonlocal ghost field predicts in this case that the outcomes on the two stern - gerlach detectors is the same with a 75% confidence while hidden variables will lead to @xmath131 confidence . the quantum correlations induced by the nonlocal quantum ghost ( [ qnlgf ] ) are `` stronger '' compared to the one obtained in the frame work of local objective realities ( [ lhv ] ) . the two different representations of quantum mechanics given by a local and nonpositive ghost field and by a nonlocal and positive ghost field are equivalent descriptions of spin-@xmath12 correlations and both have a hidden - variable look - a - like form . these descriptions are different in form , but are really equivalent to the standard formulation of quantum mechanics . instead of spin observables , projection operators and quantum states , the quantum ghost field description of the epr correlations involves local ( or nonlocal ) , positive ( or nonpositive ) distributions for random spin orientations satisfying ( or not satisfying ) the malus law . the relations ( [ qlgf ] ) and ( [ qnlgf ] ) show that the violation of bell s inequality can be due either to a nonpositive and local or to a positive and nonlocal distribution function . we have exhibited in this paper the two extreme cases which show that quantum mechanics is equivalent to a hidden - variable theory with nonpositive probabilities or to a hidden - variable theory with nonlocal distribution functions . which view we adopt is quite irrelevant because the two pictures are equivalent and represent simply different aspects of the same * quantum mechanical reality*. the author is indebted to prof . p. l. knight for the encouragement to write in this form my thoughts about the ghost field , which i presented for the first time at imperial college in 1986 . over these years i have benefited enormously discussing with i. biaynicki - birula , c. caves , t. cover , j. h. eberly , g. herling , p. meystre , g. milburn , p. milonni , k. rzaewski , m. o. scully , j. tke and a. zeilinger . krzysztof wdkiewicz is professor at the institute of theoretical physics of the university of warsaw in poland and a research professor at the center for advanced studies and the department of physics of the university of new mexico , usa . his scientific interests include quantum optics , radiation theory , stochastic processes , foundations of quantum mechanics .
einstein podolsky rosen quantum correlations are discussed from the perspective of a ghost field introduced by einstein . the concepts of ghost field , hidden variables , local reality and the bell inequality are reviewed . in the framework of the correlated singlet state of two spin-@xmath0 particles , it is shown that quantum mechanics can be cast in a way that has the form of either a nonpositive and local ghost field or a positive and nonlocal ghost field .
over the past decade it has become clear that radio sources emanating from super - massive black - holes in the centres of galaxies are - rather than merely a by - product of galaxy evolution - key players in the formation and evolution of galaxies throughout the universe . it is now fairly well established that the properties of active galactic nuclei ( agn ) are intimately linked with the properties of their host galaxies . while empirical evidence has established the co - evolution of massive black holes and their host galaxies at various redshifts ( e.g. @xcite , @xcite , see also @xcite ) , numerical simulations demonstrate that agn - induced outflows are important feedback processes for clearing the circum - nuclear regions and halting the growth of the super - massive black - holes , regulating the correlations found between the black - hole mass and the host galaxy s bulge properties and preventing the formation of too many massive galaxies in the early universe ( e.g. @xcite , @xcite , @xcite , @xcite ) . an important phase of agn feedback is when synchrotron jets emanating from the central black - hole region start propagating into the host galaxy s inter - stellar medium ( ism ) . at high redshifts , radio sources are observed to be aligned with optical continuum from warm gas and stars ( @xcite , @xcite ) and the gas in these regions is characterised by highly disturbed kinematics ( @xcite , @xcite , @xcite ) . this indicates that radio sources vigorously interact with the host galaxy s ism , entrain and expel gas and induce or quench star formation ( e.g. * ? ? ? * ) . at intermediate and low redshifts , powerful radio jets are know to induce fast ( @xmath2 km s@xmath3 ) outflows of ionised and neutral gas ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) and can shock - heat the molecular gas @xcite , exerting significant feedback on the host galaxy s ism . while propagating radio jets thus have a profound impact on their surrounding ism , the reverse is also true ( see simulations by e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? it is well established that external factors , in particular the inter - stellar and inter - galactic medium ( ism and igm ) , shape the morphological and physical properties of radio sources . for example , there is strong evidence that radio sources in general show a degree of intrinsic asymmetry ( apart from orientation effects ) , which is most prominent in compact sources @xcite . compact radio sources also show a larger degree of polarisation asymmetry than extended sources , suggesting a link between the propagating radio jets and surrounding ism @xcite . also , the impact of supersonic jets onto the ism / igm can create cocoons of ( shocked ) gas and synchrotron emission with ` hot - spot ' morphologies , while entrainment of gas by slower jets can lead to bright radio sources with edge - darkened lobes ( see e.g. reviews by * ? ? ? * ; * ? ? ? * ; * ? ? ? these external factors may thus contribute to the classical dichotomy in radio sources between fanaroff @xmath0 riley type - i ( fr - i ; * ? ? ? * ) objects , with low - power , sub - relativistic jets that have an edge - darkened morphology , and type - ii ( fr - ii ) sources , with powerful , relativistic jets ending in a bright hot - spot ( see e.g. * ? ? ? a particularly interesting object in this respect is the nearby ( @xmath4 ) radio galaxy ngc 3801 . ngc 3801 is classified as an fr - i radio galaxy , based on its total radio power ( @xmath5whz@xmath3 ; * ? ? ? * ) , radio source morphology @xcite and weak emission - line spectrum @xcite . the radio source of ngc 3801 is small ( @xmath6 kpc ; * ? ? ? * ) and contained well within the optical boundaries of a peculiar early - type host galaxy , with a patchy dust feature along the major stellar axis and a perpendicular prominent inner dust - lane @xcite , as well as a distorted and box - shaped faint outer envelope @xcite . the inner dust - lane coincides with a kpc - scale co disc , oriented perpendicular to the radio axis ( @xcite , see also @xcite ) . x - ray observations by @xcite showed evidence for shocked x - ray gas in the central region of ngc 3801 , created by supersonic , over - pressured jets that are not in equilibrium with the hot ism . these x - ray studies also revealed a high absorbing column towards the nucleus of ngc 3801 , likely due to an obscuring torus @xcite . in addition , single - dish observations showed that ngc 3801 is rich in neutral hydrogen ( hi ) gas @xcite . based on its cold gas content and peculiar optical host galaxy morphology , it has been suggested that ngc 3801 could have been involved in a gas - rich galaxy - galaxy merger @xcite . as speculated by @xcite , these are all properties that are generally not associated with fr - i sources , but are believed to be common among more powerful , high - excitation radio - loud agn ( see also @xcite , @xcite , @xcite , @xcite , as well as sect . [ sec : classical ] ) . in this paper we present radio synthesis observations of the hi gas in ngc 3801 . our observations reveal that the hi gas is distributed in a large - scale disc of cold gas . the large - scale hi disc is aligned with the radio axis , indicating that the radio jets propagate directly into the cold gas disc . based on our hi results , we discuss the formation history of ngc 3801 and investigate further how unique ngc 3801 is compared to other nearby fr - i radio galaxies . throughout this paper we assume h@xmath7 km s@xmath3mpc@xmath3 , which puts ngc 3801 at a distance of 48.6 mpc and 1 arcsec = 0.24 kpc . hi observations of ngc 3801 were performed with the westerbork synthesis radio telescope ( wsrt ) on 13 july 2009 during a single 12h track . a standard calibration was applied , using 3c147 as primary calibrator . data were reduced and analysed using the miriad and karma software . after calibration , the continuum ( ` channel-0 ' ) data were separated from the line data by fitting a straight line to the line - free channels . a radio continuum map with uniform weighting and beam - size @xmath8 ( pa -0.4@xmath9 ) was produced by fourier transforming the channel-0 data and subsequently cleaning the signal . in the same way , an hi cube with robust weighting + 1 @xcite , beam - size of @xmath10 ( pa 0.2@xmath9 ) and channel separation of 4.1 km s@xmath3 was created from the line - data . we subsequently hanning smoothed the line - data to an effective velocity resolution of @xmath11 km s@xmath3 , resulting in a noise level of 0.7 mjybm@xmath3chan@xmath3 . total intensity images of the hi emission were made by using a mask ( derived by smoothing the positive signal both spatially and kinematically ) to extract the hi line emission from the original line data - set . the data in this paper are corrected for primary beam attenuation and presented in optical barycentric velocity definition . figure [ fig : hidisc ] show the large - scale hi disc of ngc 3801 , which has a diameter of 34 kpc . a small fraction of the hi emission - line gas in the disc was detected with vla observations by @xcite . we derive a total mass of the emission - line gas of @xmath1 , though we note that part of the hi disc in our wsrt observations is detected in absorption against the radio continuum source ( see fig . [ fig : absorption ] ) and not taken into account in this hi mass estimate . the hi emission - line gas is aligned with the host galaxy s main stellar axis and the east - west dust - lane . the hi disc shows regular rotation with a large velocity spread of @xmath12 km s@xmath3 , corresponding to a maximum rotational velocity of @xmath13 km s@xmath3 ( which is larger than the rotational velocity of the optical emission - line gas ; * ? ? ? * ; * ? ? ? the hi disc is centred on v@xmath14 km s@xmath3 ( @xmath15 ) , which we argue reflects the systemic velocity of ngc 3801 . despite the regular rotation , fig . [ fig : hichanmaps ] shows that the hi disc is not ( yet ) fully settled . the most prominent asymmetry in the hi distribution occurs at the north - western edge of the disc , with gas apparently stretching in the direction of the close gas - rich companion ngc 3802 . the close companion ngc 3802 is a disc - galaxy ( s0 or spiral ) approximately 33 kpc north of ngc 3801 . ngc 3802 contains @xmath16 of hi gas . the hi gas has an asymmetric distribution across ngc 3802 , peaking in the western part of the disc and continuing on the eastern side , covering a total velocity range of @xmath17400 km s@xmath3 ( see fig . [ fig : hidisc ] _ - right _ for more details ) . while the total intensity image of fig . [ fig : hidisc ] _ ( left ) _ shows a spurious hi bridge in between ngc 3801 and the _ eastern _ part of the disc of ngc 3802 , we note that this is merely a projection effect . in fact , figs . [ fig : hidisc ] _ ( right ) _ and [ fig : hichanmaps ] show that there is an apparent bridge or warp of hi gas stretching from the north - western part of the hi disc of ngc 3801 towards the _ western _ part of the disc of ngc 3802 . this hi feature is too faint to be identified in the total intensity image of fig . [ fig : hidisc ] _ ( left)_. the asymmetric hi distribution in companion galaxy ngc 3802 and the apparent hi bridge in between ngc 3801 and ngc 3802 suggests that the ngc 3801/3802 system is in ongoing interaction . figure [ fig : absorption ] shows the hi absorption detected against the radio continuum source in the central part of ngc 3801 . for the radio continuum we derive a total integrated flux of @xmath18jy , corresponding to @xmath5whz@xmath3 . this is in agreement with earlier results by @xcite and @xcite . taking into account that the hi absorption spectra shown in fig . [ fig : absorption ] are not entirely mutually independent , there are three distinctive features visible . these features are visualised in fig . [ fig : fit ] and summarised in table [ tab : absorption ] . the first is a deep and narrow hi absorption at @xmath19 km s@xmath3 ( @xmath20 ) , which peaks at the location of the nucleus . the central velocity of this feature is in good agreement with the central velocity of @xmath21 km s@xmath3 that @xcite derive from optical spectroscopy . the second is a broad component , covering the central region and eastern radio lobe . figure [ fig : hidisc ] _ ( middle ) _ suggests that this broad absorption component traces hi gas as part of the large - scale hi disc , at locations where the gas is in front of the radio source . could be slightly steeper than that of the hi emission . this may occur because the spatial resolution of our observations is coarser than the underlying radio continuum structure , hence the hi absorption is less affected by beam - dilution than hi emission and could trace the more inner regions of the hi disc . ] this also means that the eastern radio jet is most likely the ` far ' side of the radio source ( where the radio continuum is behind the absorbing gas in the disc ) , while the western jet is the ` near ' side . the third feature is a narrow absorption against the western jet , which is blueshifted by @xmath1750 km s@xmath3 with respect to v@xmath22 . [ fig : hidisc ] _ ( middle ) _ shows that also this hi component is most likely part of the large - scale , edge - on hi disc that embeds the small radio source , although it can not be ruled out that this blueshifted hi absorption represents an outflow of neutral gas driven by the propagating radio jets . figure [ fig : absorption ] also shows indications of a narrow and much weaker hi absorption component against the eastern radio jet in ngc 3801 , around v@xmath173550 km s@xmath3 . higher resolution hi absorption observations are essential for studying the nature of the various absorption features in more detail . for this , we refer to ongoing work by @xcite . figure [ fig : hienvironment ] shows the hi environment of ngc 3801 , with five other galaxies detected in hi emission . table [ tab : companions ] summarises the hi properties of these systems . the bulk of the hi in these companion galaxies is distributed in disc - like structures . we also find two ` clouds ' of hi gas that are not associated with an optical counterpart in the sdss image of fig . [ fig : hienvironment ] . one of them , cloud a , has a velocity that coincides with that of the eastern part of the hi disc of ngc 3801 , as well as companion 3 , and is located 83 kpc south of ngc 3801 ( note that the velocity of cloud a is significantly offset from that of the much nearer companion 4 ; see fig . [ fig : hienvironment ] _ left _ ) . cloud b coincides in both velocity and location with the hi - rich companion 6 ( ngc 3806 ) and is most likely gas that is either stripped off , or being accreted onto , the hi disc of ngc 3806 . if these clouds of hi gas are tidal in origin , and given that the total hi mass of both clouds ( @xmath23 m@xmath24 ) is similar to that of ngc 3801 s close companion ngc 3802 , this indicates that past as well as ongoing interactions could have a profound influence on the evolution of this group of galaxies . [ tab : absorption ] .hi absorption properties [ cols="^,^,^,^,^,^ " , ] based on its peculiar optical morphology and co content in the inner few kpc , it has been suggested that ngc 3801 could have been involved in a gas - rich galaxy - galaxy merger ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? our hi results are in agreement with such a scenario , given that the formation of a large - scale hi disc can be the natural outcome of a galaxy merger over significant time - scales ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? simulations of merging galaxies by @xcite show that gaseous and stellar discs in the progenitor galaxies can be tidally disrupted , depending on the orbital parameters of the merging galaxies ( in particular direct encounters can create large gaseous tidal - tails ) . @xcite shows that part of this expelled material can be re - accreted back onto the newly formed host galaxy and settle into a large - scale gas disc on time - scales of one to several gyr . the bulk of the hi gas in the disc of ngc 3801 shows regular rotation , which means that the gas must have had enough time to settle into the observed disc - like structure . as a rough estimate of the merger time - scale , we assume that this process must have lasted at least one orbital period , or @xmath25 yr . the age of the radio source is significantly less than this ( @xmath26 yr ; * ? ? ? while there is no evidence for a direct causal connection between the possible merger event and the triggering of the radio source , there would have been a significant time - delay between the start of the merger and the onset of the current episode of the radio - agn activity . similar time - delays between merger ( and associated starburst episode ) and radio - agn activity have been observed in other nearby radio galaxies ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? we note , however , that our hi results alone do not provide conclusive evidence that the large - scale hi disc in ngc 3801 was formed as a result of a major merger . minor mergers or interactions with satellite galaxies ( similar to the ongoing interaction between ngc 3801 and ngc 3802 ) may also have played a role in both shaping the hi disc as well as triggering nuclear activity in ngc 3801 . given that the bulk of the hi gas that we find in our data cube is associated with companion galaxies that are all within @xmath27 km s@xmath3 from the central velocity of ngc 3801 , it is possible that the hi gas in the disc of ngc 3801 was delivered by former gas - rich companions over the course of several gyr ( similar to what is discussed by @xcite for the large hi disc in radio galaxy ngc1167 ) . another possibility is that accretion of cold gas played a major role in the formation of the hi disc . simulations by @xcite show that extended rings of gas in polar - ring galaxies may form naturally through the accretion of cold gas falling in along large filamentary structures , while @xcite argue that it is more likely for polar - ring galaxies to acquire their gaseous discs through tidal accretion from a gas - rich donor galaxy rather than a violent merger . a similar mechanism may have formed the large - scale hi disc in ngc 3801 . as mentioned above , the alignment of the small radio source in ngc 3801 with the large - scale outer hi disc makes ngc 3801 an ideal nearby radio galaxy for studying the early stage of radio - source evolution and its interaction with the surrounding cold ism . @xcite find evidence for shocked heated shells of hot gas associated with the radio source in ngc 3801 , caused by supersonic , over - pressured radio jets . their findings indicate that the radio source in ngc 3801 is heavily interacting with the surrounding ism . in proceedings based on high - resolution vla hi observations , @xcite report the presence of a broad , faint , blueshifted absorption and an hi clump associated with the shocked shell around the eastern lobe , possibly reflecting a gas outflow . our lower resolution hi data do not show clear evidence for a broad , blueshifted absorption , but such a faint hi absorption feature could be filled in with hi emission within the relatively large beam of our observations . we do detect an apparent blueshifted narrow absorption component against the western radio lobe , but as discussed in sect . [ sec : absorption ] , this most likely represents hi gas as part of the large - scale hi disc . similar to ngc 3801 , there are other known radio galaxies in which the radio source propagates into an outer hi disc . this includes coma a , where the radio lobes ionise part of the disc @xcite and ic 5063 , where the radio jets produce fast outflows of hi gas @xcite . particularly interesting is also the similarity with the disc - dominated radio galaxy b2 0722 + 30 , which hosts a small fr - i radio source ( similar to ngc 3801 ) that propagates close to the plane of the disc , in the direction of a heavily interacting hi - rich galaxy pair @xcite . in general , roughly 10@xmath28 of early - type galaxies outside clusters contain large amounts of hi gas ( m@xmath29m@xmath24 ) , often distributed in large - scale disc- or ring - like structures @xcite . in @xcite , we showed that the overall hi properties of nearby , non - cluster low - power radio galaxies are similar to those of radio - quiet early - type galaxies , with large - scale hi discs detected in a significant fraction of radio - loud systems . in addition , we showed that there is a clear segregation in hi content with radio size , in the sense that massive , large - scale hi discs are only found in the host galaxies of compact radio sources , while none of the more extended fr - i radio sources studied contain similar amounts of hi gas @xcite . the radio source in ngc 3801 is @xmath6 kpc in diameter @xcite . as illustrated in fig . [ fig : himasssize ] , the hi properties of ngc 3801 are thus similar to those of a significant fraction of nearby low - power compact radio sources . it is likely that in these systems the early evolution of the radio source is intimately linked with the surrounding cold ism , and that either confinement by the dense ism or inefficient fuelling ( through clumpiness of the cold ism and associated non - steady accretion ) keeps the radio source compact ( see discussions in * ? ? ? * ; * ? ? ? an alternative scenario , recently raised by @xcite in a study of the powerful compact radio source pks 1814 - 637 , suggests that the radio emission in these hi - rich compact radio sources may be boosted , at least temporarily , by the interaction of the radio jets with the rich ism , implying that intrinsically these compact sources may reflect a link to radio - agn in seyfert galaxies ( see also @xcite , where we speculate that the transition from radio - agn in seyferts to the brighter compact radio sources to classical fr - is may perhaps be linked to a transition from well - defined star - forming hi discs to low surface brightness hi discs to hi - poor host galaxies ) . ngc 3801 is classified as an fr - i radio galaxy , based on its total radio power ( @xmath5whz@xmath3 ) , radio source morphology @xcite and weak emission - line spectrum @xcite . however , x - ray observations by @xcite show that ngc 3801 has supersonic , over - pressured radio jets , not commonly found in fr - i sources @xcite . these x - ray studies also reveal a high absorbing column towards the nucleus ( possibly representing an obscuring torus ) , something that is generally not associated with fr - i sources @xcite . @xcite suggest that these properties may more closely resemble those of high - power , high - excitation fr - ii sources , which are believed to be post - merger systems that are fuelled by cold gas driven into the central region to form a radiatively - efficient classical accretion disc ( contrary to low - power / low - excitation fr - is , which are believed to be fuelled mainly through a quasi - spherical bondi accretion of circum - galactic hot gas directly onto the nucleus ; see e.g. @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . this scenario is in agreement with our hi results , which show that ngc 3801 is likely a post - merger system with large amounts of cold gas . ) , but note that the accretion rate necessary to sustain agn activity could be so low that it is well beyond our observing capabilities @xcite . ] however , as we saw in sect . [ sec : general ] , the presence of a large - scale hi disc is not unusual for low - power compact radio sources like ngc 3801 . we speculate that the possible link to radio - loud seyferts ( discussed in sect . [ sec : general ] ) could perhaps also explain some of the unusual characteristics of ngc 3801 . many seyfert - agn also show heavy obscuration in x - rays ( e.g. @xcite , @xcite , @xcite ) and can be found in disc - dominated , interacting galaxies that contain large amounts of hi gas ( e.g. * ? ? ? * ) . regardless of the exact nature of the radio - loud agn , the combined x - ray and hi results suggests that ngc 3801 is a system with over - pressured radio jets that are embedded in ( and interacting with ) a rich , ambient cold ism that has been deposited after a gas - rich merger ; features not or no longer present in more evolved fr - i radio sources . as discussed by @xcite , this resembles the case of centaurus a ( see also @xcite ) . x - ray observations of other hi - rich compact fr - i radio sources may reveal how unique ngc 3801 ( and cen a ) are in the evolution of classical radio sources . we have shown that the hi gas in ngc 3801 detected in earlier single - dish observations by @xcite and @xcite is distributed in a regularly rotating large - scale hi disc . we also detect hi gas in the close companion ngc 3802 ( which is in apparent ongoing interaction with ngc 3801 ) as well as several other galaxies and gas clouds in the immediate environment of ngc 3801 . the small radio source in ngc 3801 is propagating directly into the large - scale hi disc , making ngc 3801 an important system for investigating in detail the evolution of a small radio source through its host galaxy s cold ism . because of the similarity between the hi properties of ngc 3801 and those of other nearby low - power compact radio sources , studying ngc 3801 may reveal important insight into the early evolution of at least a significant fraction of low - power radio sources . we thank the referee dhruba saikia for suggestions that significantly improved this paper . the westerbork synthesis radio telescope is operated by the astron ( netherlands institute for radio astronomy ) with support from the netherlands foundation for scientific research ( nwo ) . a. , lim j. , ohyama y. , saikia d. j. , dihn - v - trung , croston j. h. , 2009 , in astronomical society of the pacific conference series , vol . 407 , the low - frequency radio universe , d. j. saikia , d. a. green , y. gupta , & t. venturi , ed . ,
we present observations of a large - scale disc of neutral hydrogen ( hi ) in the nearby fanaroff @xmath0 riley type - i radio galaxy ngc 3801 with the westerbork synthesis radio telescope . the hi disc ( 34 kpc in diameter and with @xmath1 ) is aligned with the radio jet axis . this makes ngc 3801 an ideal system for investigating the evolution of a small radio source through its host galaxy s cold ism . the large - scale hi disc is perpendicular to a known inner co disc and dust - lane . we argue that the formation history of the large - scale hi disc is in agreement with earlier speculation that ngc 3801 was involved in a past gas - rich galaxy - galaxy merger ( although other formation histories are discussed ) . the fact that ngc 3801 is located in an environment of several hi - rich companions , and shows indications of ongoing interaction with the nearby companion ngc 3802 , strengthens this possibility . the large amounts of ambient cold ism , combined with x - ray results by @xcite on the presence of over - pressured radio jets and evidence for an obscuring torus , are properties that are generally not , or no longer , associated with more evolved fr - i radio sources . we do show , however , that the hi properties of ngc 3801 are comparable to those of a significant fraction of nearby low - power compact radio sources , suggesting that studies of ngc 3801 may reveal important insight into a more general phase in the evolution of at least a significant fraction of nearby radio galaxies . [ firstpage ] galaxies : active galaxies : evolution galaxies : individual : ngc3801 galaxies : interactions galaxies : ism galaxies : jets
due to its large angular size , the galactic supermassive black hole candidate sgr a * is a promising laboratory for precision black hole astrophysics using very long baseline interferometry ( vlbi ) at millimeter wavelengths . previous measurements @xcite at 7 and 3.5 mm found a small instrinsic source size for sgr a * ( 100200 @xmath13as fwhm ) , but at those wavelengths interstellar scattering is the dominant contribution to the observed size . recent measurements by @xcite at 1.3 mm are the first at short enough wavelengths to avoid contamination due to interstellar scattering and at long enough baselines to achieve event - horizon scale resolution . these measurements have been used by @xcite to constrain the black hole and accretion flow parameters in terms of a 1.5d , nonrelativistic , time - steady radiatively inefficient accretion flow model ( riaf ; @xcite ) . sgr a * is also known to exhibit multiwavelength , polarized flares ( e.g. , @xcite ) . this activity has been modeled as hotspots orbiting in the inner radii @xcite , adiabatically expanding blobs @xcite and emission from a short , mildly relativistic jet @xcite . as a complementary approach to the relatively simple models outlined above , we perform relativistic radiative transfer on data from a recent 3d , general relativistic mhd ( grmhd ) simulation of a black hole accretion disk @xcite , and compare simulated synchrotron images to the vlbi observations . in addition , we produce light curves at 1.3 mm as a first step towards explaining sgr a * s flaring activity in terms of magnetic turbulence . in contrast to the results from riaf modeling , we find that for the single black hole spin considered ( @xmath0 ) , _ all _ inclinations provide excellent fits to the vlbi data . visibility profiles from low ( face - on ) inclinations ( @xmath1 ) resemble ring " models ( e.g. , @xcite ) , whereas those at higher ( edge - on ) inclinations are more gaussian . ongoing work with increased array sensitivity should be able to distinguish between these two types of visibility profiles . detection of a ring - like profile would provide an important constraint not only on the geometry of sgr a * , but also on the necessary complexity of accurate accretion flow models . these results also demonstrate the power of mm vlbi to distinguish between accretion models , as well as the model - dependence of attempts to extract black hole parameters . @xcite presented global grmhd simulations of black hole accretion disks . they used a spherical grid with polar kerr - schild coordinates , and the main untilted simulation had an effective resolution of @xmath14 , except near the poles which are underresolved . the torus was initialized with the analytic , time - steady , axisymmetric solution of @xcite and threaded with a weak poloidal magnetic field with minimum @xmath15 initially . the magnetorotational instability ( mri ) arose naturally from the initial conditions , and the disk quickly became fully turbulent . the simulation was evolved for @xmath16 m , or @xmath1740 orbits at @xmath18 m in units with @xmath19 . only data from the final @xmath20 of the simulation are used in this letter , once the disk is fully turbulent and transient effects from the initial conditions have died down . this simulation evolved an internal energy equation , and injected entropy at shocks . such a formulation does not conserve energy , and produces a more slender , cooler torus than conservative simulations which capture the heat from numerical dissipation of magnetic fields @xcite . we performed relativistic radiative transfer on the simulation data via ray tracing . starting from an observer s camera , rays are traced backwards in time assuming they are null geodesics ( geometric optics approximation ) , using the public code described in @xcite . in the region where rays intersect the accretion flow , the radiative transfer equation is solved along the geodesic @xcite in the form given in @xcite , which then represents a pixel of the image . this procedure is repeated for many rays to produce an image , and at many time steps of the simulation to produce time - dependent images ( movies ) . light curves are computed by integrating over the individual images . to calculate fluid properties at each point on a ray , the spacetime coordinates of the geodesic are transformed from boyer - lindquist to the kerr - schild coordinates used in the simulation . since the accretion flow is dynamic , light travel time delays along the geodesic are taken into account . data from the sixteen nearest zone centers ( eight on the simulation grid over two time steps ) were interpolated to each point on the geodesic . between levels of resolution near the poles , data from the higher resolution layer were averaged to create synthetic lower resolution points , which were then interpolated . little emission originates in the underresolved regions of the simulation . in this work , we consider synchrotron emission from sgr a * using the formulae from @xcite and neglecting polarization . the simulation provided mass density , pressure , velocity and magnetic field in code units . these were converted into cgs units following the procedure described in @xcite , except that we scaled the torus mass to match the observed flux of sgr a * at 1.3 mm . the length- and time - scales were set by the black hole mass , taken to be @xmath21 , at a distance of @xmath22 kpc . the temperature was taken to be thermal , with the ion and electron temperatures equal . many authors have used two temperature models of sgr a * , with weak coupling between the dynamically important ions and the cooler electrons . however , our single temperature model fits the observed flux from sgr a * at reasonable accretion rates @xmath23 and produces a spectrum near millimeter wavelengths that is consistent with observations . in this model , the gas is optically thin except at high inclinations during flares . the gas temperature is @xmath24 in the regions of peak synchrotron emissivity , which is consistent with peak electron temperatures in riaf models . we also followed @xcite and produced images with @xmath25 , keeping the total gas pressure fixed . these models became optically thick before reaching the observed 1.3 mm flux , and were inconsistent with the location of the peak in the quiescent spectrum of sgr a*. our fitting procedure closely mirrors that described in @xcite . simulated images were produced for a range of inclination angles , time - averaged accretion rates , and observer times ( sets of simulation time steps ) . the images were then averaged over 10 minute observation periods ( 8 simulation time steps ) to match the integration times used in @xcite , padded sufficiently to resolve the shortest baselines , and fourier transformed to 2d visibilities . the visibilities were rotated to a range of different orientations , @xmath26 , the projection of the black hole spin axis on the sky measured e of n. interstellar scattering was then applied by multiplication by an elliptical gaussian as described in @xcite using the fits from @xcite . sample images and scatter - broadened visibilities are shown in fig . [ imgs ] . at low inclinations , the emission region is essentially a ring around the shadow , which causes a minimum in the 2d visibilities . for higher inclinations @xmath27 , doppler beaming causes the emission to be concentrated on one side of the black hole . the visibilities are then approximately elliptical , with the minor ( major ) axis corresponding to the major ( minor ) axis of the emission region . visibilities were interpolated to the detections in @xcite , and an effective @xmath28 was computed as defined in @xcite . grids of minimum effective , reduced @xmath28 are shown in fig . [ chi2 ] over inclination angle and sky orientation . at inclinations @xmath29 , our results are in at least qualitative agreement with @xcite . the orientation is well constrained at high inclinations , since the general image shape is nearly static and asymmetric , and the long baseline vlbi measurements pick out a specific orientation for the visibility ellipse . the two distinct bands of good fits are due to the ( approximate ) up - down symmetry of the image . however , riaf models ruled out inclinations @xmath1 , whereas for our grmhd models , good fits ( reduced @xmath30 ) are possible at all inclinations . this is especially evident from the curve of @xmath31(@xmath32 ) vs. @xmath32 overplotted in fig . inclinations @xmath33 are less probable due to the @xmath34 prior and large variation of probability density with observer time . although the best fit @xmath28 is roughly the same for low inclinations at all @xmath26 , at any given observer time good fits are restricted to less than half the range of sky orientations . visibility profiles from the best fits at many inclinations are shown in fig . [ profs ] . the profiles represent the 2d visibility plotted along the line representing the average angle in the uv plane of the long baseline measurements from @xcite . also plotted are the data from that work . at low inclinations , the profile reaches a minimum as described above , whereas at @xmath29 , the profile decreases monotonically with baseline length . our best fit face - on visibility profile is almost identical to that of the ring model from @xcite . our images are strongly peaked at the circular photon orbit , as can be seen in fig . [ imgs ] . low inclination ( @xmath1 ) images produced from a single time step of the energy conserving simulation by @xcite are similar to that in fig . [ imgs ] , suggesting that our results are not due to the nonconservative nature of the simulation . it is not surprising that the vlbi observations are well fit by face - on disks . in addition to the ring model , a delta function intensity profile at the circular photon orbit roughly matches the observations , as does the spherical accretion model from @xcite scaled to the mass of sgr a*. in fact , any model sharply peaked at the circular photon orbit should approximately match the visibility data . this is because the ratio of the zeroth to first maximum in the visibility for such a model is roughly the same as that of the zeroth order bessel function , and agrees with the ratio of the visibility amplitude at @xmath35 to the total flux . if this grmhd simulation approximates the correct model for sgr a * , we can also constrain the accretion rate for the single spin value used . upper and lower limits for @xmath36 and @xmath7 confidence from probability distributions marginalized over observer time and sky orientation are shown in fig . [ mdot ] . doppler beaming leads to larger fluxes at high inclinations , and hence a lower accretion rate is necessary to match the observed flux . the limits are also more stringent at high inclinations , where the flux is less variable ( see fig . [ lcurve ] ) . further marginalizing over inclination , we constrain the accretion rate to be ( @xmath4@xmath5)@xmath37 at @xmath7 confidence . this constraint on the accretion rate of sgr a * , while model - dependent , only uses the vlbi visibility measurements . a major advantage of using grmhd simulations in place of riaf models is the ability to probe variability from the same dynamical model used to produce spectra and visibilities . previous work has used hotspots @xcite to study variability , but this introduces extra free parameters , complicating attempts to extract black hole parameters . the light curves of our sgr a * models at 1.3 mm shown in fig . [ lcurve ] are the first from simulation data to consider synchrotron emission , and the first to include the time delay along geodesics . the variability on short timescales is slightly more noticeable for the @xmath38 case . interestingly , though , the longer time scale flaring behavior is more pronounced in the @xmath39 light curve . the power spectra are well described by red noise spectra with power law index @xmath40 , @xmath41 for @xmath39 , @xmath42 . these are both steeper than the observed power law index of @xmath43 in @xcite at 3 mm wavelength . the twin flares occur simultaneously in both light curves , rise over half an hour and last @xmath9@xmath10 hours . the flux modulation is @xmath12 ( @xmath44 ) at @xmath39 ( @xmath42 ) , measured from the peak of the flare to the average of the light curve immediately preceding it . all of these features are consistent with mm flare observations of sgr a * @xcite . since the flares are seen at low inclinations , they are not caused by doppler shifts from hotspots . the peak intensity is attenuated at large inclination due to rising optical depth . the flares are caused by a rise in magnetic field strength near the midplane in the inner radii ( @xmath45@xmath46 m ) . they are not due to heat from magnetic dissipation since this is not possible in the simulation used here . thus heating from magnetic reconnection is not necessary to produce the mm flares of sgr a*. in this letter , simulated images of a single temperature synchrotron emission model from 3d , grmhd simulations by @xcite have been fit to recent 1.3 mm vlbi observations of sgr a*. for the single black hole spin considered , the high inclination results are in at least qualitative agreement with those presented for 1.5d , stationary , non - relativistic models used by @xcite . however , we find that in contrast to previous work , images at low inclination provide excellent fits to the data , which are relatively insensitive to sky orientation , and present a qualitatively different visibility profile due to approximate circular symmetry . radial intensity profiles from face - on images of recent riaf models @xcite appear more extended than those from our grmhd simulations . the minimum of the visibility profile for these images occurs at too short a baseline to match the vlbi data . the model in @xcite , on the other hand , has increasing number density up to the event horizon , in stark contrast to the results from the simulations , where the number density peaks at @xmath47 , well outside of the event horizon at @xmath48 and the marginally stable orbit at @xmath49 . large number densities in the inner radii of the model from @xcite cause the shadow " to be much smaller than expected @xcite , weakening the power on small scales ( large baselines ) , and preventing a good fit to the vlbi data at low inclinations . increased 1.3 mm array sensitivity should favor either the low or high inclination profiles in fig . [ profs ] . our predicted profiles would also be sensitive to measurements at 1.3 mm along baselines of length @xmath2@xmath3 km . a future detection favoring a ring - like profile would constitute the first measurement of a black hole shadow . the visibility analysis constrains the average accretion rate of the @xmath0 simulation to be ( @xmath4@xmath5)@xmath6 with @xmath7 confidence . while this constraint is model - dependent , it is independent of and consistent with constraints from spectral fitting @xcite and polarization measurements @xcite , and in good agreement with the previous spectral fits of single temperature grmhd synchrotron images by @xcite . we have also produced the first light curves of synchrotron emission from grmhd modeling of sgr a*. at 1.3 mm , flares arise naturally from magnetic turbulence , and are consistent with observations . unfortunately our analysis is poorly suited for studying variability in other wavebands such as the radio , nir or x - ray . in the radio , most the emission originates outside of the initial torus pressure maximum at @xmath50 , while the nir and x - ray luminosity is produced by some combination of synchrotron emission from non - thermal electrons and compton scattering of synchrotron photons . modeling of both is left for future work . in addition , polarization measurements from flares are an important observational clue . adding this capability to our relativistic radiative transfer code is in progress . there are many uncertainties in the analysis presented in this letter . while grmhd allows for a self - consistent solution of the primitive variables , it is unknown whether the mhd approximation is valid for the low densities at the galactic center . in addition , there are a range of possible solutions , depending on the energy prescription used . current simulations also neglect the interplay between ions and electrons , which should be handled self - consistently . if sgr a * is radiatively inefficient , simulations which conserve total energy are probably more appropriate . we estimate the time - averaged energy dissipation rate from numerical losses in the simulation as @xmath51 ergs @xmath52 for an accretion rate of @xmath53 . this is larger than the bolometric luminosity of sgr a * , and gives an effective radiative efficiency of @xmath54 . the accretion flow is advection dominated , in that an order of magnitude more energy is lost to the black hole than to numerical reconnection . face on synchrotron images from a single timestep of the 3d , conservative code in @xcite were similar to those found here , suggesting that our results may not be specific to nonconservative simulations . despite the uncertainties , relativistic radiative transfer is an important tool for connecting state of the art , 3d , global grmhd simulations with observations . our analysis shows that ongoing millimeter vlbi observations will not only constrain black hole parameters , but accretion models as well . expanding this analysis to the full range of spins will help determine the model dependence of such future constraints . if sgr a * is at a low inclination , mm vlbi may soon provide the first direct evidence of a black hole shadow . in addition to the visibility fits and light curves considered here , the variability of sgr a * can be probed through measurements of closure phases along triangles of baselines ( e.g. , @xcite ) . identifying salient features of these closure phases is a goal of future work . if the flares seen here are robust features of mm synchrotron light curves of grmhd simulations , multiwavelength and polarization measurements can be used to constrain emission mechanisms and the non - thermal electron distribution . thanks shep doeleman for the vlbi data , and omer blaes for useful discussions . this work was partially supported by nasa grants 05-atp05 - 96 and nnx08ax59h , and a graduate fellowship at the kavli institute of theoretical physics at the university of california , santa barbara under nsf grant phy05 - 51164 .
the recent vlbi observation of the galactic center black hole candidate sgr a * at 1.3 mm shows source structure on event - horizon scales . this detection enables a direct comparison of the emission region with models of the accretion flow onto the black hole . we present the first results from time - dependent radiative transfer of general relativistic mhd simulation data , and compare simulated synchrotron images at black hole spin @xmath0 with the vlbi measurements . after tuning the accretion rate to match the millimeter flux , we find excellent agreement between predicted and observed visibilities , even when viewed face - on ( @xmath1 ) . vlbi measurements on @xmath2@xmath3 km baselines should constrain the inclination . the data constrain the accretion rate to be ( @xmath4@xmath5)@xmath6 with @xmath7 confidence , consistent with but independent of prior estimates derived from spectroscopic and polarimetric measurements . finally , we compute light curves , which show that magnetic turbulence can directly produce flaring events with @xmath8 hour rise times , @xmath9@xmath10 hour durations and @xmath11@xmath12 flux modulation , in agreement with observations of sgr a * at millimeter wavelengths .
obesity is determined through the body mass index ( bmi ) which compares the weight and height of an individual via the formula weight(kg)/height(m@xmath0 ) . a bmi value of 30 is considered the obesity threshold . overweight but not obese is @xmath51bmi@xmath52 , and underweight is [email protected] . our main measure in this work is the adult obesity prevalence of a county , @xmath54 , for a given year defined as the number of obese adults @xmath31 ( bmi@xmath55 ) in a county @xmath4 over the total number of adults in this county , @xmath12 . we use the data from the usa center for disease control ( cdc ) downloaded from @xcite . cdc provides an estimate of the obesity country - wide since 1970 , at the state level from 1984 to 2009 , and at the county level from 2004 to 2008 . the study of the correlation function @xmath2 requires high resolution data . therefore , we use data defined at the county level and restrict our study of obesity and diabetes to the available period 2004 - 2008 . other indicators are provided by different agencies at the county level for longer periods . the datasets analyzed in this paper were obtained from the websites as indicated below . they can be downloaded as a single tar datafile from jamlab.org . the datasets consist of a list of populations and other indicators at specific counties in the usa at a given year . a graphical representation of the obesity data can be seen in fig . [ transition]b for usa from 2004 to 2008 , where each point in the maps represents a data point of obesity prevalence directly extracted from the dataset . the datasets that we use in our study have been collected from the following sources : * ( a ) population. * us census bureau . we download a number of datasets at the county level from http://www.census.gov/support/usacdatadownloads.html - for the population estimates we use the table pin030 . for the years 1969 - 2000 we use data supplied by bea ( bureau of economic analysis ) and for years 2000 - 2009 we use the file co-est2009-alldata.csv from http://www.census.gov/popest/counties/files/co-est2009-alldata.csv * ( b ) health indicators. * data downloaded from the centers for disease control and prevention ( cdc ) . the center provides county estimates between the years 2004 - 2008 for : - diagnosed diabetes in adults . - obesity prevalence in adults . - physical inactivity in adults . the estimates for obesity and diabetes prevalence and leisure - time physical inactivity were derived by the cdc using data from the census and the behavioral risk factor surveillance system ( brfss ) for 2004 , 2005 , 2006 , 2007 and 2008 . brfss is an ongoing , state - based , random - digit - dialed telephone survey of the u.s . civilian , non - institutionalized population aged 18 years and older . the analysis provided by the brfss is based on self - reported data , and estimates are age - adjusted on the basis of the 2000 us standard population . full information about the methodology can be obtained at http://www.cdc.gov/diabetes/statistics . * ( c ) economic indicators. * we download data for economic activity through http://www.census.gov / econ/. the economic activity of each sector is measured as the total number of employees in this sector per county in a given year normalized by the population of the county . the north american industry classification system ( naics ) ( http://www.census.gov/epcd/naics02/naicod02.htm ) assigns hierarchically a number based on the particular economy sector . the naics is the standard used by us statistical agencies in classifying business establishments across the us business economy . in this study we have used the following economic sectors with their corresponding naics : * whole economy . entire output of all economic sectors combined including all naics codes . * \31 . manufacturing . broad economic sector from textiles , to construction materials , iron , machines , etc . * \42 . wholesale trade . very broad sector including merchants wholesalers , motors , furniture , durable goods , etc . * \56 . administrative jobs and support services . * \445 . food and beverage stores . including all the food sectors , from supermarkets , fish , vegetables meat markets , to restaurants and bars and other services to the food industry . supermarkets and other grocery ( except convenience ) stores . this is a subsection of naics 445 . food services and drinking places . a sub - sector of naics 72 which includes restaurants , cafeterias , snacks and nonalcoholic beverage bars , caterers , bars and drinking places ( alcoholic beverages ) . * ( d ) mortality rates. * we use data from the national cancer institute seer , surveillance epidemiology and end results downloaded from http://seer.cancer.gov/data/ the institute provides mortality data from 1970 to 2003 , aggregated every three years . we analyze the mortality of a specific form of cancer per county normalized by the population of the county . here , we use mortality data for the following causes of death : - all cancer , independently of type . - lung cancer . we take advantage of the available data of population distribution around the globe defined in a square grid of 2.5 arc - seconds obtained from @xcite . this data allows to study the correlation functions of the population distribution for many countries . by using this data we are able to test the system size dependence of our results we find that the correlation length @xmath15 is proportional to the linear size of the country , @xmath25 . the linear size is calculated as total area = @xmath56 . we find that the correlation scales with the system size as discussed in the text . for instance , for the usa population distribution we find @xmath22 km , while a smaller country like uk has @xmath57 km . table [ xi ] shows a list of countries used in figs . [ correl]b and c to determine the correlation length @xmath15 of the correlation function of population density . the shape of the main obesity clusters and location of the red bonds and obesity epicenter are depicted in fig . [ percolation]c overlayed with a us map showing the boundaries of states and counties . figure [ percolation]c shows the obesity clusters obtained at @xmath58 , @xmath43 , @xmath59 , and @xmath46 , depicting the process of percolation . at @xmath58 , we plot the largest red cluster which is seen in the lower mississippi basin . the highest obesity prevalence is in greene county , al , which acts as the epicenter of the epidemic . at @xmath44 , we plot in yellow the second largest cluster in the atlantic region south of the appalachian mountains , and at @xmath59 we plot the third largest cluster ( violet ) , which appears north of the appalachian mountains . we mark with black the three red bonds that make the mississippi cluster to grow abruptly by absorbing the clusters in the appalachian range . the red bonds are dekalb county , tn , mclean county , ky , and colquitt county , ga . this transition is reflected in the jump in the size of the largest cluster in fig . [ percolation]a . the same process is observed in the second percolation transition at @xmath60 , when the red bond , rich county , ut , joins the eastern and western clusters for a whole - country percolation . the scaling properties characterizing the geometry and distribution of clusters at percolation are @xcite : _ ( i ) _ the scaling of the number of boxes @xmath61 to cover the infinite spanning cluster versus the size of the boxes @xmath62 : @xmath63 defining the fractal dimension of the spanning cluster , @xmath47 . _ ( ii ) _ the number of boxes , @xmath64 , of size @xmath62 covering the perimeter of the infinite cluster : @xmath65 defining the hull fractal dimension , @xmath48 . _ ( iii ) _ the probability distribution of the area of clusters at percolation : @xmath66 characterized by the critical exponent @xmath49 . additionally , there is a scaling relation between the fractal dimension and the cluster distribution exponent @xcite : @xmath67 the exponents @xmath68 for percolation with long - range correlations have been calculated numerically in @xcite as a function of the correlation exponent @xmath17 using standard percolation analysis . there exist also a theoretical prediction based on renormalization group in @xcite for the correlation length exponent . the values of @xmath68 for the obesity clusters at the first percolation transition , @xmath44 , are reported in the main text . a direct computer simulations of long - range percolation @xcite for @xmath33 finds the values of the three geometric exponents to be @xmath69 , consistent with those reported here . we notice that the exponent @xmath49 is expected to be larger than 2 . this is due to mass conversation , assuming that the power - law eq . ( [ tau ] ) extends to infinity at percolation in a infinite system size . the fact that we find a value slightly smaller than 2 for the obesity clusters , might be due to a finite size effect . we also notice that the values of the exponents obtained from correlated percolation at @xmath33 are not too far from those of uncorrelated percolation @xcite . therefore , the values of the exponents may not be enough to precisely compare the obesity clusters with long - range percolation clusters . however , they serve as an indication that the obesity clusters have the geometrical properties of clusters at a critical point , such as scaling behavior . furthermore , it could be possible that long - range correlated percolation may capture only part of the dynamics of the clustering epidemic . it could be , for instance , that higher order correlations , beyond the two - point correlation captured by @xmath2 , are also relevant in determining the value of the exponents . in this case , our analysis should be supplemented by studies of @xmath70point correlation functions , beyond @xmath2 . .list of countries used to calculate the correlation length , @xmath15 , from the correlation function of the population density . data is obtained from @xcite . @xmath25 is calculated as the square root of the total area of the country . [ cols="^,>,>",options="header " , ] our approach supplements covariance analysis @xcite . instead , we use physics concepts to shed a different view on the spreading of epidemics . our approach can be extended to the study of the geographical spreading of any epidemic : from diabetes and lung cancer , to the spreading of viruses or real states bubbles , where the spatial spreading plays an important role . population correlations are naturally inherited by all demographic observables . even variables whose incidence varies randomly from county to county would exhibit spatial correlations in their absolute values , simply because its number increases in more populated counties and population locations are correlated . indeed , the absolute number of obese adults per county is directly proportional to the population of the county [ bettencourt , l. m. a. , _ et al . _ growth , innovation , scaling , and the pace of life in cities . usa _ * 104 * , 7301 - 7306 ( 2007 ) ] . our aim is to measure spatial fluctuations on the frequency of incidence , independent of population agglomeration . thus , spatial correlations of all indicators ought to be calculated on the density defined , in the case of obesity , as @xmath71 rather than on the absolute number of obese people , @xmath31 , itself . the spatial correlations of the fluctuations of @xmath40 from the global average captures the collective behavior expressed in the power - law described in eq . ( [ gamma ] ) .
non - communicable diseases like diabetes , obesity and certain forms of cancer have been increasing in many countries at alarming levels . a difficulty in the conception of policies to reverse these trends is the identification of the drivers behind the global epidemics . here , we implement a spatial spreading analysis to investigate whether non - communicable diseases like diabetes , obesity and cancer show spatial correlations revealing the effect of collective and global factors acting above individual choices . specifically , we adapt a theoretical framework for critical physical systems displaying collective behavior to decipher the laws of spatial spreading of diseases . we find a regularity in the spatial fluctuations of their prevalence revealed by a pattern of scale - free long - range correlations . the fluctuations are anomalous , deviating in a fundamental way from the weaker correlations found in the underlying population distribution . the resulting scaling exponents allow us to broadly classify the indicators into two universality classes , weakly or strongly correlated . this collective behavior indicates that the spreading dynamics of obesity , diabetes and some forms of cancer like lung cancer are analogous to a critical point of fluctuations , just as a physical system in a second - order phase transition . according to this notion , individual interactions and habits may have negligible influence in shaping the global patterns of spreading . thus , obesity turns out to be a global problem where local details are of little importance . interestingly , we find the same critical fluctuations in obesity and diabetes , and in the activities of economic sectors associated with food production such as supermarkets , food and beverage stores which cluster in a different universality class than other generic sectors of the economy . these results motivate future interventions to investigate the causality of this relation providing guidance for the implementation of preventive health policies . the world health organization has recognized obesity as a global epidemic @xcite . obesity heads the list of non - communicable diseases ( ncd ) like diabetes and cancer , for which no prevention strategy has managed to control their spreading @xcite . here , since the gain of excessive body weight is related to an increase in calories intake and physical inactivity @xcite a principal aspect of prevention has been directed to individual habits @xcite . however , the prevalence of ncds shows strong spatial clustering @xcite . furthermore , obesity spreading has shown high susceptibility to social pressure @xcite and global economic drivers @xcite . this suggests that the spread and growth of obesity and other ncds may be governed by collective behavior acting over and above individual factors such as genetics and personal choices @xcite . to study the emergence of collective dynamics in the spatial spreading of obesity and other ncds , we implement a statistical clustering analysis based on critical phenomenon physics . we start by investigating regularities in obesity spreading derived from correlation patterns of demographic variables . obesity is determined through the body mass index ( bmi ) obtained via the formula weight(kg)/[height ( m)]@xmath0 . the obesity prevalence is defined as the percentage of adults aged @xmath1 18 years with a bmi @xmath1 30 . we investigate the spatial correlations of obesity prevalence in the usa during a specific year using micro - data defined at the county - level provided by the us centers for disease control ( cdc ) @xcite through the behavioral risk factor surveillance system ( brfss ) from 2004 to 2008 ( see methods section [ data ] ) . the average percentage of obesity in usa was historically around 10% . in the early 80s , an obesity transition in the hitherto robust percentage , steeply increased the obesity prevalence ( fig . [ transition]a ) . the spatial map of obesity prevalence in the usa shows that neighboring areas tend to present similar percentages of obese population @xcite forming spatial ` obesity clusters ' @xcite . the evolution of the spatial map of obesity from 2004 to 2008 at the county level ( fig . [ transition]b ) highlights the mechanism of cluster growth . characterizing such geographical spreading presents a challenge to current theoretical physics frameworks of cluster dynamics @xcite . the equal - time two - point correlation function , @xmath2 , determines the properties of such spatial arrangement by measuring the influence of an observable @xmath3 in county @xmath4 ( e.g. , in this study : adult population density , prevalence of obesity and diabetes , cancer mortality rates and economic activity ) on another county at distance @xmath5 @xcite : @xmath6 here , @xmath7 is the average over all @xmath8 counties in the usa , @xmath9 is the standard deviation , @xmath10 is the euclidean distance between the geometrical center of counties @xmath4 and @xmath11 . large positive values of @xmath2 reveal strong correlations , while negative values imply anti - correlations , i.e. , two areas with opposed tendencies relative to the mean in obesity prevalence ( analogous to two domains with opposite spins in a ferromagnet @xcite ) . spatial correlations in any indicator ought to be referred to the natural correlations of population fluctuations ( fig . [ transition]c ) . to this aim , we first calculate @xmath2 for the population ( adults @xmath1 18 years ) in usa counties , @xmath12 , by using the density : @xmath13 in eq . ( [ cr ] ) , where @xmath14 is the county area . population density correlations show a slow fall - off with distance ( fig . [ correl]a ) described by a power - law up to a correlation length @xmath15 : @xmath16 where @xmath17 is the correlation exponent . correlations become short - ranged when @xmath18 ( @xmath19 is the dimension of the map ) , and stronger as @xmath17 decreases @xcite . an ordinary least squares ( ols ) regression analysis @xcite on the population reveals the exponent @xmath20 ( averaged over 1969 - 2009 , fig . [ correl]a , error bars denote 95% confidence interval [ ci ] ) . the inset of fig . [ correl]a reveals a distance where correlations vanish , @xmath21 with @xmath22 km , representing the average size of the correlated domains @xcite . as we increase @xmath5 larger than @xmath15 , we consider correlations between areas in the east and west which are anti - correlated since @xmath23 for @xmath24 . to determine whether population correlations are scale - free , we calculate @xmath2 for geographical systems of different sizes using a high resolution grid of 2.5 arc - seconds , available for several countries from @xcite ( methods section [ gridded ] ) . the resulting correlations ( fig . [ correl]b ) reveal the same picture as for the usa at the county - level ( fig . [ correl]a ) , i.e. , a power - law up to a correlation length . we then measure @xmath15 for every country , and investigate whether , as expected with the laws of critical phenomena @xcite , it increases with the country size , @xmath25 . indeed , we obtain ( fig . [ correl]c and table [ xi ] ) , @xmath26 where @xmath27 is the correlation length exponent @xcite . this result implies that the fluctuations in human agglomerations are scale - free , i.e. , the only length - scale in the system is set by its size and the correlation length become infinite when @xmath28 @xcite . we interpret any departure from @xmath29 as a proxy of anomalous dynamics beyond the simple dynamics related to the population growth . when we calculate the spatial correlations of obesity prevalence ( @xmath30 , @xmath31 is the number of obese adults in county @xmath4 ) in usa from 2004 to 2008 we also find long - range correlations ( fig . [ correl]a ) . the crux of the matter is that the correlation exponent for obesity ( @xmath32 , averaged over 2004 - 2008 ) is smaller than that of the population , signaling anomalous growth . since smaller exponents mean stronger correlations , the increase in obesity prevalence in a given place can eventually spread significantly further than expected from the population dynamics . we also calculate fluctuations in variables which are known to be strongly related to obesity @xcite : diabetes and physical inactivity prevalence ( fraction of adults per county who report no physical activity or exercise , see methods section [ data ] ) . the obtained @xmath17 exponents are anomalous with similar values as in obesity ( fig . [ time]a ) . the system size dependence of @xmath15 for obesity and diabetes can not be measured directly , since there is no available micro - data for other countries . however , we find that @xmath15 of obesity and diabetes in usa is very close to @xmath15 of the population distribution ( inset of fig . [ correl]a ) . assuming that the equality of the correlation lengths holds also for other countries , then obesity and diabetes should satisfy eq . ( [ sf ] ) as well . the correlations in obesity are reminiscent of those in physical systems at a critical point of second - order phase transitions @xcite . physical systems away from criticality are uncorrelated and fluctuations in observables , e.g. , magnetization in a ferromagnet or density in a fluid , decay faster than a power - law , e.g. , exponentially @xcite . instead , long - range correlations appear at critical points of phase transitions where fluctuations are not independent and , as a consequence , fall - off more slowly . the existence of long - range correlations with @xmath33 rather than the noncritical exponential decay signals the emergence of strong critical fluctuations in obesity and diabetes spreading . the notion of criticality , initially developed for equilibrium systems @xcite , has been successfully extended to explain a wide variety of dynamics away from equilibrium ranging from collective behaviour of bird cohorts , biological and social systems to city growth , just to name a few @xcite . its most important consequence is that it characterizes a system for which local details of interactions have a negligible influence in the global dynamics @xcite . following this framework , the clustering patterns of obesity are interpreted as the result of collective behaviors which are not merely the consequence of fluctuations of individual habits . as a tentative way of addressing which elements of the economy may be related to the obesity spread , we calculate @xmath17 in economic indicators which are thought to be involved in the rise in obesity @xcite . except for transient phenomena , all studied indicators yield exponents that fall around @xmath34 or @xmath35 , representing two universality classes of weak and strong correlations , respectively ( figs . [ time]a and b ) . we begin by studying the correlations in the whole economy ( measured through the number of employees of all economic sectors per county population , see methods section [ data ] ) . we find @xmath17 close to @xmath36 ( over the period 1998 - 2009 , fig . [ time]b and c ) suggesting that the whole economy inherits the correlations in the population . generic sectors of the economy which are not believed to be drivers of obesity , e.g. , wholesalers , administration , and manufacturing , also display @xmath17 consistent with the population trend ( fig . [ time]b and c ) . interestingly , analysis of the spatial fluctuations in the economic activity of sectors associated to food production and sales ( supermarkets , food and beverages stores and food services such as restaurants and bars ) gives rise to the same anomalous value as obesity and diabetes ( @xmath37 , 1998 - 2009 , fig . [ time]b and c ) . although these results can not inform about the causality of these relations , they show that the scaling properties of the obesity patterns display a spatial coupling which is also expressed by the fluctuations of sectors of the economy related to food production . it is of interest to study other health indicators for which active health policies have been devoted to control the rate of growth . we apply the correlation analysis to lung cancer mortality defined at the county level and compare with cancer mortality due to all types ( methods section [ data ] ) . the spatial correlations of cancer mortality per county show an interesting transition in the late 70 s from anomalous strong correlations , @xmath38 , to weak correlations , @xmath36 , ( fig . [ time]a and d ) . this transition is visualized in the different correlated patterns of cancer mortality in 1970 and 2003 in fig . [ transition]d , i.e. , the clustering of the data is more profound in 1970 , while in 2003 it spreads more uniformly . the current status of all - cancer mortality fluctuations is close to the natural one , inflicted by population correlation . conversely , fluctuations in the mortality rate due to lung cancer from 1970 to 2003 have remained highly correlated and close to the obesity value , @xmath38 ( fig . [ time]a and [ transition]e ) , while the other types of cancer have become less correlated . this is an interesting finding since , similarly to obesity , lung cancer prevalence is affected by a global factor ( smoking ) and has been growing rapidly during the studied period . the most visible characteristic of correlations is the formation of spatial clusters of obesity prevalence . to quantitatively determine the geographical formation of obesity clusters , we implement a percolation analysis @xcite . the control parameter of the analysis is the obesity threshold , @xmath39 . an obesity cluster is a maximally connected set of counties for which @xmath40 exceeds a given threshold @xmath39 : @xmath41 . by decreasing @xmath39 , we monitor the progressive formation , growth and merging of obesity clusters . in random uncorrelated percolation @xcite , small clusters would be formed in a spatially uniform way until a critical value , @xmath42 , is reached , and an incipient cluster spans the entire system . instead , when we analyze the obesity clusters we observe a more complex pattern exemplified in fig . [ percolation]a and [ percolation]b for year 2008 . at large @xmath39 , the first cluster appears in the lower mississippi basin ( red in fig . [ percolation]a ) with epicenter in greene county , al . upon decreasing @xmath39 to 0.32 , new clusters are born including two spanning the south and north of the appalachian mountains , which acts as a geographical barrier separating the second and third largest clusters ( yellow and violet in fig . [ percolation]a , respectively ) . further lowering @xmath39 , we observe a percolation transition in which the appalachian clusters merge with the mississippi cluster . this point is revealed by a jump in the size of the largest component and a peak in the second largest component at @xmath43 ( fig . [ percolation]a ) as features of a percolation transition @xcite . at @xmath44 , three `` red bonds '' ( mclean county , ky , dekalb county , tn , and colquitt county , ga ) appear to connect the incipient largest cluster spanning the east of usa ( see fig . [ percolation]c and methods section [ red ] ) . as a comparison , when we randomize the obesity data by shuffling the values between counties , a single @xmath45 appears as a signature of a uncorrelated percolation process ( blue symbols in fig . [ percolation]a ) . obesity clusters in the west persist segregated from the main eastern cluster avoiding a full - country percolation due to low - prevalence areas around colorado state . finally , the east and west clusters merge at @xmath46 by a red bond ( rich county , utah ) producing a second percolation transition ; this time spanning the whole country ( see fig . [ percolation]a and c , where the whole spanning cluster is green , and methods section [ red ] ) . this cluster - merging process is a hierarchical percolation progression represented in the tree model in fig . [ percolation]b . to further inquire whether the spreading of obesity has the features of a physical system at the critical point , we examine the geometry and distribution of obesity clusters . for long - range correlated critical systems percolating through nearest neighbors in two dimensional maps , the geometrical structure @xcite gives rise to three critical exponents : the fractal dimension of the spanning cluster , @xmath47 , the fractal dimension of the hull , @xmath48 , and the cluster size distribution exponent , @xmath49 , analogous to zipf s law @xcite ( methods section [ som - percolation ] ) . for the percolating obesity cluster at @xmath44 displayed in the inset of fig . [ percolation]d , we confirm critical scaling with exponents : @xmath50 ( fig . [ percolation]d , e ) . taken together , these results show that obesity spreading behaves as a self - similar strongly - correlated critical system @xcite . in particular , a note of caution has to be raised since , even if the highest prevalence of obesity is localized to the south and appalachia , the scaling analysis indicates that the obesity problem is the same ( self - similar ) across all usa , including the lower prevalence areas . our results can not establish a causal relation between obesity prevalence and economic indicators : whether fluctuations in the food economy may impact obesity or , instead , whether the food industry reacts to obesity demands . however , the comparative similarities of statistical properties of demographic and economical variables serves to identify possible candidates which shape the epidemic . specifically , the observation of a common universality class in the fluctuations of obesity prevalence and economic activity of supermarkets , food stores and food services which cluster in a different universality class than simple population dynamics is in line with studies proposing that an important component of the rise of obesity is linked to the obesogenic environment @xcite regulated by food market economies @xcite . this result is consistent with recent research that relates obesity with residential proximity to fast - food stores and restaurants @xcite . the present analysis based on clustering and critical fluctuations is a supplement to studies of association between people s bmi and food s environment based on covariance @xcite ( methods section [ covariance ] ) . we have detected potential candidates in the economy which relate to the spreading of obesity by showing the same universal fluctuation properties . these tentative relations ought to be corroborated by future intervention studies . 99 world health organization . obesity : preventing and managing the global epidemic . _ who obesity technical report series _ * 894 * ( world health organization geneva , switzerland , 2000 ) . butland , b. _ et al . _ , foresight tackling obesities : future choices - project report , 2nd edn . london : government office for science ( 2007 ) . hill , j. o. & peters j. c. environmental contributions to the obesity epidemic . _ science _ * 280 * , 1371 - 1374 ( 1998 ) . swinburn , b. a. , sacks , g. , hall , k. d. , mcpherson , k. , finegood , d. t. , moodie , m. l. & gortmaker , s. l. the global obesity pandemic : shaped by global drivers and local environments . _ lancet _ * 378 * , 804 - 814 ( 2011 ) . nestle , m. _ food politics : how the food industry influences nutrition and health . _ ( university of california press , revised edition , 2007 ) . christakis , n. a. & fowler , j. h. the spread of obesity in a large social network over 32 years . _ n. engl . j. med . _ * 357 * , 370 - 379 ( 2007 ) . block , j. , christakis , n.a . , omalley , a.j . & subramanian , s. proximity to food establishments and body mass index in the framingham heart study offspring cohort over 30 years . _ american journal of epidemiology _ * 174 * , 1108 - 1114 ( 2011 ) . caballero , b. & popkin , b. m. , editors . _ the nutrition transition : diet and disease in the developing world _ ( academic press , 2002 ) . cutler , d. m. , glaeser , e. l. & shapiro j. m. why have americans become more obese ? _ j. econ . perspect . _ * 17 * , 93 - 118 ( 2003 ) . haslam d. w. & james , w. p. obesity . _ lancet _ * 366 * , 1197 - 1209 ( 2005 ) . department of health and human services ( usa ) . healthy people 2010 : understanding and improving health . conference edition . ( washington , government printing office , 2000 ) . schuurman , n. , peters , p. a. & oliver , l. n. are obesity and physical activity clustered ? a spatial analysis linked to residential density . _ obesity _ * 17 * , 2202 - 2209 ( 2009 ) . michimi , a. & wimberly , m. c. spatial patterns of obesity and associated risk factors in the conterminous u.s . _ am . j. prev . med . _ * 39 * , e1-e12 ( 2010 ) . behavioral risk factor surveillance system survey data . atlanta , georgia : u.s . department of health and human services , centers for disease control and prevention . http://apps.nccd.cdc.gov/ddt_strs2/nationaldiabetesprevalenceestimates.aspx stanley , h. e. _ introduction to phase transitions and critical phenomena . _ ( oxford university press , oxford , 1971 ) . bunde , a. & havlin , s. ( editors ) . _ fractals and disordered systems _ ( springer - verlag , heidelberg , 2nd edition , 1996 ) . coniglio , a. , nappi , c. , russo , l. & peruggi , f. _ j. phys . a _ * 10 * , 205 - 209 ( 1977 ) weinrib , a. long - range correlated percolation . _ phys . rev . b _ * 29 * , 387 - 395 ( 1984 ) . prakash , s. , havlin , s. , schwartz , m. & stanley , h. e. structural and dynamical properties of long - range correlated percolation . _ phys . rev . a _ * 46 * , r1724 - 1727 ( 1992 ) . makse , h. a. , havlin , s. & stanley , h. e. modelling urban growth patterns . _ nature _ * 377 * , 608 - 612 ( 1995 ) . mora , t. & bialek , w. are biological systems poised at criticality ? _ j. stat . phys _ * 144 * , 268 - 302 ( 2011 ) . montogomery , d. c. , peck , e. a. introduction to linear regression analysis ( wiley , new york , 1992 ) . cavagna , a. , cimarelli , a. , giardina , i. , parisi , g. , santagati , r. stefanini , f. & viale , m. scale - free correlations in starling flocks . _ proc . natl . acad . sci . usa _ * 107 * , 11865 - 11870 ( 2010 ) . center for international earth science information network ( ciesin ) , columbia university ; and centro internacional de agricultura tropical ( ciat ) . 2005 . gridded population of the world , version 3 ( gpwv3 ) . palisades , ny : socioeconomic data and applications center ( sedac ) , columbia university . available at http://sedac.ciesin.columbia.edu/gpw . mokdad , a. h. _ et al . _ , prevalence of obesity , diabetes , and obesity - related health risk factors , 2001 . _ jama _ * 289 * , 76 - 79 ( 2003 ) . rozenfeld , h. d. , rybski , d. , gabaix , x. & makse , h. a. the area and population of cities : new insights from a different perspective on cities . _ american economic review _ * 101 * , 560 - 580 ( 2011 ) . stanley , h. e. , _ et al . _ scale - invariant correlations in the biological and social sciences . _ phil . mag . part b _ * 77 * , 1373 - 1388 ( 1998 ) . weinrib , a. & halperin , b. i. critical phenomena in systems with long - range - correlated quenched disorder . _ phys . rev . b _ * 27 * , 413 - 427 ( 1983 ) . swinburn , b. & figger , g. preventive strategies against weight gain and obesity . _ obesity reviews _ * 3 * , 289 - 301 ( 2002 ) . chang , v. w. & christakis , n.a . income inequality and weight status in u.s . metropolitan areas . _ social science and medicine _ * 61 * , 83 - 96 ( 2005 ) . * acknowledgements : * lkg and ham are supported by nsf-0827508 emerging frontiers , and pb and ms by human frontiers science program . we are grateful to h. nickerson for bringing our attention to this problem , and s. alarcn - daz and d. rybski for valuable discussions . we thank epiwork , arl and the israel science foundation for support . * methods *
photons are an excellent probe of the hot and dense , strongly interacting matter produced in heavy ion collisions @xcite . they do not participate in the strong interaction and thus exit the system carrying information from the time of their emission , allowing a glimpse at the time - evolution of the matter . experimentally we measure a time - integrated history of the emission . photons from hadron decays need to be removed to reveal the so - called direct contribution , i.e. photons that are produced before the formation of the matter as well as from the matter itself . further removal of the early component , usually considered prompt production from @xmath8 scattering of the partons from the incoming nuclei , gives access to the radiation emitted from the matter . if the matter is in local equilibrium the photon spectrum is a time - integrated image of the evolution of the temperature and collective motion of the matter as it expands and cools . phenix discovered evidence of thermal photons from collisions at the relativistic heavy ion collider ( rhic ) @xcite ; similar findings have recently been reported by alice from collisions at the large hadron collider @xcite . photons in both energy regimes exhibit a large yield and an azimuthal anisotropy @xcite with respect to the reaction plane , often referred to as elliptic flow and quantified as @xmath9 . comparing the measured spectra to model calculations of thermal photons based on a hydrodynamic evolution of the system , microscopic transport models , or a more schematic time evolution gives reasonable agreement when assuming an initial temperature of 300 mev or above @xcite for @xmath10 gev @xmath11 collisions at rhic . however , it is a challenge for these types of models to simultaneously explain the large observed azimuthal anisotropy of the radiation and the large yield @xcite . the challenge for these model calculations results from the interplay between the time evolution of the collective motion and the cooling of the matter that emits photons . in the model calculations , the collective motion builds up over time . the flow velocity is initially small and increases throughout the collision as the matter continues to expand . the yield of thermal photons is expected to be largest early in the collision when the matter is the hottest . theoretical models that create large photon @xmath9 typically underestimate the direct photon yield . attempts to improve hydrodynamic models by implementing next - to - leading - order thermal rates @xcite , initial state fluctuations @xcite , formation time effects @xcite , increased radial flow and enhanced coupling at @xmath12 @xcite , fail to reconcile yield and anisotropy . to resolve this puzzle , new production mechanisms have been proposed . some enhance the thermal yield in the presence of the strong magnetic field perpendicular to the reaction plane , which creates a large anisotropy @xcite . other new mechanisms , such as synchrotron radiation @xcite at the plasma boundary or photon production in a glasma phase @xcite , create an anisotropy due to the initial geometry of the overlap region . in this paper we present the first measurement of low momentum real direct photons from collisions at 200 gev center of mass energy . this measurement compliments earlier measurements of direct photons that were obtained by extrapolating low mass virtual photons to the real photon point @xcite . we are able to extend the range down to 0.4 gev / c and provide new information on the centrality dependence of the direct photon yield . in particular , the centrality dependence holds the promise to help to distinguish between different production mechanisms @xcite . to measure direct photons , we analyzed large data samples of @xmath13 and @xmath14 minimum - bias au@xmath15au collisions recorded with the phenix central arm spectrometers during the 2007 and 2010 runs , respectively . the main phenix detector is described in detail elsewhere @xcite . in addition , a hadron blind detector ( hbd ) @xcite was installed , except for part of the 2007 rhic run when only one half of the hbd was installed . the data were taken with a special field configuration which essentially cancels the magnetic field around the beam axis out to about @xmath16 . minimum - bias events were triggered using the beam - beam counters ( bbc ) that cover the rapidity region @xmath17 and @xmath18 in azimuth in both beam directions . the bbc information is used to limit the vertex in beam direction to @xmath19 around the nominal position . the charge measured in the bbc is used to categorize the event centrality . the sample is divided into four centrality classes , 0%20% for the most central selection , 20%40% , 40%60% and 60%92% for the most peripheral sample . the raw inclusive photon yield is measured through photon conversions to pairs in the detector material , which allows us to avoid hadron contamination and measure photons down to @xmath20 . trajectories and momenta of @xmath21 and @xmath22 are determined using the drift chambers and the pad chambers that measure the deflection in the axial magnetic field together with the interaction vertex location . we require a minimum of @xmath23 . the energy is determined with the electromagnetic calorimeter ( emcal ) . the @xmath21 and @xmath22 are identified utilizing the ring - imaging erenkov detector by requiring a minimum of three phototubes associated with both charged tracks at the expected ring radius as well as requiring the respective energy / momentum ratios to be greater than 0.6 . we select photons that converted in the readout plane of the hbd that is located at a radius of @xmath24 and has a radiation length @xmath25 . our method to identify photon conversions uses only the phenix central arm detectors , with the hbd playing no active role . because the draft chambers are located at @xmath26220 cm radially from the beam axis , the momentum reconstruction algorithm needs to assume where the particles originate . in the standard algorithm all charged particles are reconstructed as if they came from the event vertex . this procedure mismeasures the momentum vector for @xmath21 and @xmath22 from photon conversions in the hbd . for conversions in the hbd readout plane the artificial opening angle of the pair is @xmath2610 mrad and the pair momentum increases by 1%2% . as a result the pair is reconstructed with an average mass of @xmath27 , as is shown in the invariant - pair - mass distribution of fig . [ fig:1dmassplot]a . the first peak in the mass plot at a few mev/@xmath28 is from @xmath29 dalitz decays , along with a small number of pairs from photon conversions before the hbd readout plane . the momenta of all low mass pairs are recalculated assuming that they originated at the hbd readout plane . if the pair is indeed a conversion pair from the readout plane , the relative momentum resolution of the pair is approximately @xmath30 and the pair mass recalculated with the hbd back plane as origin ( @xmath31 ) is a few mev/@xmath28 , consistent with the experimental resolution . for all other pairs , the momentum vectors are now mismeasured , in particular pairs from @xmath29 dalitz decays are now reconstructed with larger opening angles and thus shifted upward in pair mass . the recalculated mass spectrum is shown in panel ( b ) of fig . [ fig:1dmassplot ] . plotting the yield as a function of @xmath31 versus the mass calculated with the vertex as origin ( @xmath32 ) , as shown in fig . [ fig:2dmassplot ] , allows one to clearly isolate the conversions in the hbd readout plane . we select photon conversions by a two dimensional cut @xmath33 and @xmath34 , illustrated by the red dashed box . note that the large distance from the true event vertex and the relatively thick hbd readout plane ( in terms of radiation length @xmath35 ) with no comparable radiating material nearby makes identification of the converted photons very accurate : a full geant monte carlo simulation @xcite shows that the purity of this sample is 99% , with most of the remaining 1% being photon conversions at other radii . @xmath22 pair invariant - mass distribution from data . panel ( a ) shows the distribution of masses calculated with the normal reconstruction algorithm ( vtx ) . panel ( b ) shows the distribution of masses calculated with the alternate track model assumption ( hbd ) . ] a subset of this inclusive conversion photon sample is tagged statistically as photons from @xmath29 decays if they reconstruct the @xmath29 mass with a second , photon - like shower taken from the emcal . note that this is done in bins of , the transverse momentum of the converted photons , not in bins of @xmath29 . a cut on the shower shape of this second emcal shower is used to remove most hadrons . false tagging from hadron showers in the emcal is further reduced by applying a lower threshold on the cluster energy . for the 2010 data we applied an @xmath360.4 gev cut , which is just above the emcal response for minimum ionizing particles . for the 2007 data , a higher threshold of 0.6 gev was necessary due to a cut on the shower energy that was introduced during data production . invariant - mass distributions for two different bins . the left column ( a),(c ) displays the mass for @xmath37 gev/@xmath2 , the right column ( b),(d ) displays the mass for @xmath38 gev/@xmath2 . the top row ( a),(b ) shows the @xmath39 invariant - mass foreground distribution in blue , with the normalized background distribution from the mixed events in red . the bottom row ( c),(d ) shows the isolated pion peak after subtraction of the normalized background . the masses are calculated from the hbd readout plane origin assumption on the electron tracks . the centrality bin is 0%20% . ] in each bin the number of @xmath29 tagged photons ( ) is determined by integrating the @xmath40 mass distribution around the @xmath29 mass after subtraction of the mixed - event combinatorial background . figure [ fig : pi0massplot ] shows the mass distributions before and after subtracting the mixed - event background for two example bins ( 0.40.6 gev/@xmath2 and 1.82.0 gev/@xmath2 ) for central collisions ( 0%20% ) , which have the smallest signal - to - background ratio . the @xmath29 peak extraction method has less than 4% systematic uncertainty on the @xmath29 tagged photon yield , which is assumed to be independent between neighboring bins and thus folded into the statistical uncertainties . in a given bin the true number of inclusive photons @xmath41 and photons from @xmath29 decays @xmath42 are related to the measured quantities and as follows : @xmath43 @xmath44 where @xmath2 is the probability that the photon converts in the hbd readout plane , @xmath45 is the reconstruction efficiency of the pair and @xmath46 is the factor describing that both @xmath21 and @xmath22 are in the detector acceptance . the factor @xmath47 is the conditional acceptance that after one photon from a @xmath29 decay was reconstructed as conversion pair , the partner photon falls into the acceptance of the emcal . the probability that the partner photon is reconstructed is given as @xmath48 . the product @xmath49 is averaged over all possible of the partner photon , indicated by @xmath50 . because and are both measured in terms of the of the converted photon , the efficiency and acceptance factors for the pair as well as the conversion probability explicitly cancel in the ratio @xmath51 . this ratio can be converted into , the ratio of the yield of true inclusive photons @xmath41 to the yield of true photons from hadron decays @xmath52 : @xmath53 all terms in eq . [ eqn : rgamma ] are a function of converted photon . will be unity for a given bin if all photons result from hadron decays , or larger than unity if direct photons are present in the sample . the excess above unity is a measure of the direct photon content in the bin . in the following we discuss all terms in detail . decay in the emcal , if the other photon converted in the hbd electronics and was reconstructed as an pair . the abscissa gives , the of the pair . the cut of 0.6 ( 2007 ) and 0.4 gev/@xmath2 ( 2004 ) is on the photon detected in the emcal . for the cut of 0.6 gev/@xmath2 we show the results for two methods , a full mc simulation ( points ) and a fast mc simulation ( histogram ) . for the cut of 0.4 gev , the fast mc simulation is shown as dashed histogram . ] lccc source & @xmath54 & type & + @xmath29 reconstruction & & & + ( tagged photon yield ) & 4% & a & + @xmath55 purity & 1% & c & + + conditional acceptance @xmath56 & & & + energy scale & 4% & b & + conversion loss & 2% & c & + @xmath55 efficiency & 1% & b & + active area & 1% & c & + input spectra & 1% & b & + + @xmath57 & & & + @xmath58 ratio & 2.2% & c & + other mesons & @xmath591% & c & + [ tab : uncertainties ] yield extraction . they are plotted as vertical lines . all other systematic uncertainties are added in quadrature and shown as filled boxes . _ ( b ) _ in the combined 2007 + 2010 measurement . ] the numerator of eq . [ eqn : rgamma ] includes the measured ratio @xmath51 , and the efficiency and acceptance correction for pion tagging , . figure [ fig : ef ] shows for the min . bias data sets of 2007 and 2010 . it increases monotonically with and is lower for the larger cut on the second photon . these trends can be understood in terms of decay kinematics and average of the tagged @xmath29 . at higher the average of the tagged @xmath29 is larger , the opening angle between the decay photons becomes smaller and the probability to have both decay photon in the phenix acceptance increases . consequently increases with . a larger -cut on the second photon increases the minimum @xmath29 necessary for both photons to be accepted at a given , thus is larger for the lower cut . the ratio of for the two different cuts is as large as a factor of 2 at the lowest and decreases towards higher . because the final result for is proportional to , varying the cut provides a powerful cross check for the measurement . we developed two different methods to determine @xmath60 . for the 2007 data a geant monte carlo simulation of the detector response to @xmath29 decays is performed . in the simulation one photon is forced to convert in the hbd readout plane . the simulated @xmath29 decays are then embedded into real data to account for occupancy effects in the emcal . the events are analyzed through the full reconstruction chain to extract @xmath60 . this method is computationally very intensive and thus limited by statistical uncertainties . to overcome these we developed a fast simulation . it accounts for the detector acceptance and variations of the active detector areas with time . the single photon response is parametrized based on a geant monte carlo simulation of single photons . to test the fast simulation we compared its result for in fig . [ fig : ef ] for the 2007 data to the one determined with the full geant simulation ; the two methods agree within statistical uncertainties . for the 2010 data we used the fast simulation . we also compare results on for cuts on the second photon between 0.3 and 0.6 gev/@xmath28 and find that the results are consistent well within the systematic uncertainties on discussed below . the denominator of eq . [ eqn : rgamma ] is the ratio of photons from all hadron decays ( @xmath52 ) to those from @xmath29 decays ( @xmath42 ) . to evaluate this ratio , the per - event yields @xmath42 and @xmath52 are determined using the phenix meson decay generator exodus , which is discussed in detail in ref . @xcite . for each centrality class , a fit to the measured per - event yields for charged and neutral pions @xcite is used to generate @xmath29 s that then are decayed to photons according to known branching ratios and decay kinematics based on ref . the resulting photon spectrum is the per - event yield @xmath42 as a function of photon @xmath4 . to generate @xmath52 , the contributions from decays of @xmath61 , @xmath62 , and @xmath63 are determined using the same procedure and then added to @xmath42 . the shape of the spectra for @xmath61 , @xmath62 , and @xmath63 are derived from the @xmath29 spectrum by replacing with @xmath64 . for @xmath61 and @xmath62 this is consistent with published data @xcite ; for @xmath63 no data are available . the absolute normalization of the @xmath61 per - event yield is set using a value of @xmath65 @xcite at = 5 gev/@xmath2 . for the @xmath62 and @xmath63 the absolute yield is set to @xmath66 and @xmath67 , again at 5 gev/@xmath2 ( see @xcite ) . several sources contribute to systematic uncertainties on . the largest one is 4% and accounts for the uncertainties of the energy scale and the energy resolution . these translate directly into an uncertainty in the number of photons that pass the lower emcal threshold and thus become candidates for @xmath29 tagging . the second largest uncertainty ( 2% ) is on the number of photons that are lost because they convert to pairs in the detector material in front of the emcal and are not reconstructed as single showers . the active area of the detectors was studied as a function of time , and the resulting systematic uncertainty on is smaller than 1% . varying the @xmath29 input distribution with the uncertainties on the data results in a 1% uncertainty on . lastly , the uncertainty on the photon reconstruction efficiency is also small ( 1% ) , estimated by varying the shower shape cuts , redoing the analysis and recalculating the correction , and comparing the results . all other systematic effects were found to be negligible . systematic uncertainties on @xmath57 are dominated by the accuracy with which @xmath58 is known . because the @xmath29 contribution to @xmath52 is @xmath2680% , the systematic uncertainty on the @xmath29 spectra largely cancels , leaving the @xmath58 ratio as the dominant source of systematic uncertainties . the uncertainty on also includes possible deviations from scaling with @xmath68 and uncertainties on the other meson yields . the total uncertainty is less than 2.5% . all systematic uncertainties on are summarized in table [ tab : uncertainties ] . as function of photon for the combined 2007 and 2010 data sets in centrality bins 0%20% , 20%40% , 40%60% and 60%92% . statistical uncertainties plotted as vertical lines are dominated by the @xmath29 yield extraction . all other systematic uncertainties are added in quadrature and shown as filled boxes . on panels ( a ) and ( b ) we also show earlier results from ref . @xcite , obtained by extrapolating virtual photons to zero mass . ] figure [ fig : rg_mb ] compares our results for in minimum - bias collisions from the 2007 and 2010 data sets separately , while figure [ fig : rg ] shows the same quantity for the four centrality selections . here we used the full geant simulation for the 2007 data , and the fast monte carlo simulation for the 2010 data . from the two data sets agree well within statistical errors . figure [ fig : rg ] also includes data from an earlier publication @xcite , in which was obtained by extrapolating virtual photons to @xmath69 for the two central bins and @xmath70 . the was used to calculate the direct - photon spectra shown in @xcite ; here we show the corresponding data points . we observe no statistically significant difference between the measured from real and virtual photons . however , given the uncertainties , we can not rule out a difference of up to 15% , as is estimated in ref . the shows a statistically significant excess of photons above those expected from hadron decays , and this excess increases with centrality . au collisions from this measurement ( solid symbols ) and au@xmath15au and collisions ( open symbols ) . open circles and up triangles : low spectrum obtained with virtual photons in and @xcite . open squares and down triangles : spectrum of real photons , measured in the emcal in . open squares are 2003 data @xcite , open down triangles are 2006 data @xcite . open stars : spectrum with real photons , measured in the emcal in au@xmath15au in 2004 @xcite . the dashed line is a fit to the combined set of data , extrapolated below 1gev/@xmath2 , and the solid line the fit scaled with the number of minimum - bias au@xmath15au collisions . bands around lines denote @xmath71 uncertainty intervals in the parameterizations of the data and the uncertainty in , added in quadrature . ] , cf . [ fig : direct_yield_mb ] . one - sided errors denote @xmath71 upper limits , other uncertainties are as in fig . [ fig : direct_yield_mb ] . ] to combine the data sets we apply the corrections calculated from the fast simulation for both the 2007 and 2010 data ( after verifying consistency between the corrections calculated for the 2007 data with both the fast monte carlo and full geant ) and average the numerators in eq . [ eqn : rgamma ] for the 2007 and 2010 data sets . while the correction factor is different for the two data sets ( due to differences in detector dead areas and the different minimum photon energy cuts applied ) , the systematic uncertainties are the same . next we determine the direct photon yield from the combined @xmath72 for each bin : @xmath73 were @xmath52 is the invariant yield of photons from hadron decays , which we calculate from measured charged and neutral pion spectra , as described above . at this point a systematic uncertainty of 10% on the shape of the input @xmath29 distribution for the generator needs to be included @xcite ( this mostly cancels in the denominator of @xmath74 , but no longer cancels in eq . [ eqn : directyield ] ) . the measurement was cross - checked and found consistent with the direct photon spectrum calculated using the fully corrected measured inclusive photon spectrum @xcite via the relation @xmath75 , which has much larger systematic uncertainties because the conversion probability , the @xmath3 pair efficiency and acceptance do not cancel . figure [ fig : direct_yield_mb ] shows the direct photon spectra for minimum bias and our previously published data from ref . @xcite and @xcite . also shown are the photon data from phenix . the lowest points ( open circles ) come from a virtual photon measurement @xcite , while the open squares and open triangles are from the analysis of the 2003 @xcite and 2006 @xcite data sets , respectively . the dashed curve is the joint fit to the data with a functional form @xmath76 . this shape was used in ref . including new data in the fit @xcite , we find parameters @xmath77 , @xmath78 and @xmath79 . note that the systematic uncertainties are highly correlated . also , the lowest actual data point in the fit is at = 1gev/@xmath2 . the solid curve in fig . [ fig : direct_yield_mb ] is the fit scaled by the corresponding average number of binary collisions , @xmath80 , for minimum - bias collisions , as calculated from a glauber monte carlo simulation @xcite . below @xmath81 , an enhancement above the expected prompt production ( ) is observed . the enhancement has a significantly smaller inverse slope than the @xmath5 scaled contribution . figure [ fig : pt_direct ] shows that we observe similar behavior when investigating the centrality dependence in more detail . the solid curves are again the fit scaled by the respective number of binary collisions , and they deviate significantly from the measured yields below 3gev/@xmath2 . . dashed lines are fits to an exponential function in the range @xmath82 . ] . the number of nucleon participants , number of binary nucleon - nucleon collisions , and constituent - quark participants vs centrality bin . also shown are the values of local inverse slopes in the range 0.6 to 2 gev/@xmath2 of the direct photon spectra , after subtracting the scaled results . [ cols=">,^,^,^,<",options="header " , ] [ tab : powerfit ] the yield increases with a power - law function @xmath6 ; this is illustrated by the linear rise of the yield with in the logarithmic representation shown on fig . [ fig : int_yield ] together with fits to @xmath83 . the fit parameters are shown in table [ tab : powerfit ] . the same power is observed independent of the cutoff , consistent with the spectra having the same shape independent of centrality . a simultaneous fit to the data in fig . [ fig : int_yield ] results in an average value of @xmath84 . we have also considered the recently suggested scaling with the number of quark participants , which works well for charged particle production @xcite . here is calculated with a glauber monte carlo simulation similar to by picking random locations for constituent quarks within the nucleus . while our data is better described by scaling with a power - law in , it is also consistent with a power - law function @xmath85 , where @xmath86 is the number of quark participants . in this case we find an exponent of @xmath87 . in most theoretical models thermal photon emission involves binary collisions of constituents , partons or hadrons , in hot and dense matter . thus the emission rate from a unit volume should be proportional to the square of the number of constituents , while bulk particle production should scale with the number of constituents @xcite . because particle production is approximately proportional to @xmath80 one might expect thermal photon emission to scale as @xmath88 times a correction for the increasing reaction volume with centrality . the increasing volume will reduce the centrality dependence , so that one expects @xmath89 for thermal photon emission , just as observed . recent theoretical studies of the centrality dependence confirm our finding that the yield of thermal photon emission increases approximately with a power law function of . in the phsd transport approach the power @xmath90 is approximately 1.5 @xcite , with no evident change in the shape of the spectra with centrality , very similar to our data . a hydrodynamic model @xcite shows a power law increase of the yield with a power @xmath90 in the range from 1.67 to 1.9 , increasing monotonically as the lower integration threshold increases from 0.4 to 1.4 gev/@xmath2 . photon production in a glasma phase @xcite was predicted to scale with @xmath6 with @xmath91 . other new production mechanisms , proposed to address the large @xmath9 , have distinctly different centrality dependence . the yield from enhanced thermal photon emission in the strong magnetic field is expected to decrease with centrality , as the strength of the field weakens with decreasing impact parameter @xcite . the thermal photon yield should thus increase more slowly than expected from standard processes , but a quantitative estimate is not yet available . we have isolated the low momentum direct photon yield emitted in collisions . the shape of the spectra does not depend strongly on centrality , with an average inverse slope of 240 mev/@xmath2 in the range from 0.6 to 2 gev/@xmath2 . the yield increases with centrality as @xmath6 with @xmath92 . in conclusion , these results will help distinguish between different photon - production mechanisms and will constrain models of the space - time evolution of heavy ion collisions . we thank the staff of the collider - accelerator and physics departments at brookhaven national laboratory and the staff of the other phenix participating institutions for their vital contributions . we acknowledge support from the office of nuclear physics in the office of science of the department of energy , the national science foundation , abilene christian university research council , research foundation of suny , and dean of the college of arts and sciences , vanderbilt university ( u.s.a ) , ministry of education , culture , sports , science , and technology and the japan society for the promotion of science ( japan ) , conselho nacional de desenvolvimento cientfico e tecnolgico and fundao de amparo pesquisa do estado de so paulo ( brazil ) , natural science foundation of china ( p. r. china ) , croatian science foundation and ministry of science , education , and sports ( croatia ) , ministry of education , youth and sports ( czech republic ) , centre national de la recherche scientifique , commissariat lnergie atomique , and institut national de physique nuclaire et de physique des particules ( france ) , bundesministerium fr bildung und forschung , deutscher akademischer austausch dienst , and alexander von humboldt stiftung ( germany ) , hungarian national science fund , otka , kroly rbert university college ( hungary ) , department of atomic energy and department of science and technology ( india ) , israel science foundation ( israel ) , national research foundation of korea of the ministry of science , ict , and future planning ( korea ) , physics department , lahore university of management sciences ( pakistan ) , ministry of education and science , russian academy of sciences , federal agency of atomic energy ( russia ) , vr and wallenberg foundation ( sweden ) , the u.s . civilian research and development foundation for the independent states of the former soviet union , the hungarian american enterprise scholarship fund , and the us - israel binational science foundation . 42ifxundefined [ 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 * * , ( ) link:\doibase 10.1088/1742 - 6596/446/1/012028 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevc.89.034906 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevd.89.026013 [ * * , ( ) ] link:\doibase 10.1134/s0021364013150071 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2013.01.014 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.109.152302 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevc.89.034908 [ * * , ( ) ] @noop * * , ( )
the phenix experiment at rhic has measured the centrality dependence of the direct photon yield from collisions at @xmath0 gev down to @xmath1 gev/@xmath2 . photons are detected via photon conversions to @xmath3 pairs and an improved technique is applied that minimizes the systematic uncertainties that usually limit direct photon measurements , in particular at low @xmath4 . we find an excess of direct photons above the @xmath5-scaled yield measured in collisions . this excess yield is well described by an exponential distribution with an inverse slope of about 240 mev/@xmath2 in the range from 0.62.0 gev/@xmath2 . while the shape of the distribution is independent of centrality within the experimental uncertainties , the yield increases rapidly with increasing centrality , scaling approximately with @xmath6 , where @xmath7 .
a great deal of attention has been focused recently on the creation of strong nonlinear effects in single coherently prepared multi - level atoms @xcite . the motivation for this interest is to fabricate an atomic medium with giant nonlinear properties produced with relatively low light powers . a particular attention has been paid to the third - order kerr - type nonlinearities which play an important role in nonlinear optics and have many fascinating applications in different areas of physics ranging from phase modulation @xcite , generation of optical solitons @xcite , optical switching @xcite to optical communication and computing @xcite . important for practical applications is to achieve enhanced or giant kerr nonlinearities in an atomic medium with significantly reduced or even completely cancelled absorption rate for the propagating light beam . et al . _ @xcite have proposed a scheme to produce giant kerr nonlinearities together with reduced absorption , by using quantum interference effects related to electromagnetically induced transparency . in a four - level double - dark resonance system , kerr nonlinearity can be enhanced several orders of magnitude accompanied by vanishing linear absorption under the condition of the effective interaction of double dark resonances @xcite . a number of different atomic schemes have been suggested to achieve a large nonlinearity with vanishing absorption @xcite . more recently , niu and gong @xcite and yan _ et al . _ @xcite have shown that the kerr nonlinearity can be enhanced with vanishing linear and nonlinear absorptions due to the spontaneously generated coherence @xcite . the major obstacle in experimental investigations of the nonlinear properties of multi - level atoms is the difficulty to find suitable systems to create quantum interference effects between atomic transitions responsible for the cancelation of the absorption of a propagating field . most of the schemes proposed have assumed that the quantum interference occurs between two transitions with parallel or anti - parallel dipole moments . in atoms with quantum states close in energy the dipole moments are usually perpendicular . therefore , several schemes have been suggested to engineer quantum interference effects in atoms with perpendicular dipole moments . most of the schemes suggests to use single - mode optical cavities with preselected polarization in bad cavity limit @xcite . et al . _ @xcite have found that the purcell effect @xcite can substantially influence the kerr nonlinearity . et al . _ @xcite proposed a method to produce self- and cross - kerr photonic nonlinearities using light induced stark shifts arising from the interaction of a cavity mode with atoms . in addition to the purcell effect which is substantial in optical cavities , where spontaneous transitions occur only at selected frequencies , a cavity - induced quantum interference is expected to arise which is analog of the spontaneously generated interference @xcite . thus , a question arises , to what extent a combination of the purcell effect and the cavity - induced interference effects will affect the susceptibility of driven @xmath0-type three - level atom . the purpose of this paper is to address this question and discuss in detail the possibility of obtaining giant kerr nonlinearities . we consider a three - level atom in the @xmath0 configuration in which one of the two dipole allowed transitions is driven by a strong laser field while the other is probed by a weak beam . the atomic transitions are simultaneously coupled to a tunable single - mode cavity . our interest will be centered principally on the effect of the cavity on the third - order susceptibility and determine if the driven system possesses enhanced or giant nonlinearities accompanied by vanishing absorption . the paper is organized as follows . in section [ ch2 ] , we introduce our model and outline the major steps in the derivation of the equations of motion for the density matrix elements . the iterative analytical solution for the coherences determining the susceptibility is presented in section [ ch3 ] . the results are presented graphically and discussed in section [ ch4 ] . we show the influences of the cavity field frequency , the cavity field - atom coupling constants and the atomic decay constants on the real and imaginary parts of the linear and nonlinear susceptibilities . we summarize our results in section [ ch5 ] . we consider a @xmath0-type three - level atom composed of two excited states @xmath1 and @xmath2 coupled to a common ground state @xmath3 by transition dipole moments @xmath4 and @xmath5 , respectively . the atom is located inside a single - mode cavity field of frequency @xmath6 and polarization @xmath7 , as shown in fig . [ fig1 ] . . the atomic transition @xmath8 is driven by a strong laser field of frequency @xmath9 and is probed by a weak laser field of a tunable frequency @xmath10 coupled to the @xmath11 transition . each of the two laser fields couples only to one of the atomic dipole transitions , while the cavity field couples to both transitions.,width=302 ] the polarization @xmath7 is chosen such that the cavity field is simultaneously coupled to both atomic transitions with the coupling strengths @xmath12 and @xmath13 , respectively . the atomic transition @xmath14 is driven by a strong laser field of frequency @xmath9 , whereas the @xmath15 transition is probed by a weak tunable laser beam of frequency @xmath10 . the cavity mode is damped at the rate @xmath16 , whereas the atomic transitions are damped by spontaneous emission to the modes other than the cavity mode at the rates @xmath17 and @xmath18 , respectively . in a frame rotating at the frequency @xmath9 , the master equation of the density operator @xmath19 of the overall system ( the atom and the cavity field ) is of the form @xmath20+{\cal l}_{a}{\rho_{t}}+{\cal l}_{c}{\rho_{t } } , \label{e1}\end{aligned}\ ] ] where @xmath21 is the unperturbed hamiltonian of the coherently driven and weakly probed atom , @xmath22 is the hamiltonian of the cavity field , @xmath23 is the interaction hamiltonian of the cavity field with the atomic transitions , and @xmath24 are dissipative terms describing the damping of the atomic transitions by spontaneous emission and of the field by cavity decay . here , @xmath25 and @xmath26 are the annihilation and creation operators for the cavity field , @xmath27 are the atomic operators , @xmath28 is the frequency difference between the atomic transitions , @xmath29 , @xmath30 and @xmath31 are the detunings of the atomic frequency @xmath32 , the probe beam frequency @xmath10 and the cavity frequency @xmath6 from the driving laser frequency @xmath9 . the parameters , @xmath33 and @xmath34 are the ( real ) rabi frequencies of the driving laser field of amplitude @xmath35 and of the probe beam of amplitude @xmath36 . in writing the master equation ( [ e1 ] ) , we have assumed that the atomic dipole moments are not orthogonal to each other , which results in the cross damping terms between the atomic transitions . these terms lead to quantum interference between the two transitions and are determined by the so - called cross damping parameter @xmath37 , where @xmath38 is the angle between @xmath4 and @xmath5 . when the dipole moments are parallel ( @xmath38 = 0 ) , the cross damping parameter is maximal with @xmath39 , whilst @xmath40 when the dipole moments are perpendicular @xcite . quantum interference has been studied intensively over years and has revealed new phenomena of both conceptual and practical importance . it has been shown that interference between atomic transitions induced by external fields or by spontaneously created atomic coherence can lead to novel phenomena such as electromagnetically induced transparency , lasing without inversion , enhanced index of refraction and also nonlinear processes such as enhanced the kerr nonlinearities . however , most of the predicted quantum interference effects have so far eluded observation , as it is very unlikely to find isolated atoms with two non - orthogonal dipole moments and states close in energy . therefore , we propose an alternative scheme where one can engineer coherence between atomic transitions with perpendicular dipole moments by coupling the transitions to a single - mode cavity field . as we shall see , crucial for the creation of the coherence is to couple the cavity mode simultaneously to both of the atomic transitions . in practice , it can be easily achieved by setting a cavity - field polarization making , for example , an angle @xmath41 with the direction of the atomic dipole moment @xmath4 and that simultaneously forms the angle @xmath42 with the dipole moment @xmath5 . the master equation ( [ e1 ] ) we have started with is written in the basis of the atomic states . since the atomic transition @xmath43 is driven by a strong laser field , it prompts us to introduce dressed states which provide a good approach for studying the problem . the dressed states are eigenstates of the hamiltonian @xmath44 and are defined by the eigenvalue equation @xmath45 whose the eigenvalues and corresponding eigenstates , in the limit of a weak probe beam @xmath46 , are @xmath47 where @xmath48 and @xmath49 is the detuned rabi frequency of the driving field . we now introduce the interaction between the dressed atom and the cavity field and work in the bad cavity limit @xcite , in which the cavity decay dominates over the coupling strengths @xmath50 and the atomic decay rates @xmath17 and @xmath18 , i.e. @xmath51 such a feature implies that the cavity mode response to the standard vacuum reservoir is much faster than that produced by its interaction with the atom . in other words , the cavity field forms a finite bandwidth ( markovian ) vacuum reservoir . working in the bad cavity limit , we can adiabatically eliminate the cavity variables . this yields a master equation where the damping terms have a structure dependent on the difference between the cavity field and the dressed - atom transition frequencies . details of the adiabatic approximation have been presented in ref . @xcite . here , we will apply such approach to study the linear and nonlinear responses of the system to a weak probe field . from the cavity - modified master equation , the equation of motion for the atomic density matrix elements , written in the dressed state basis , are of the form @xmath52 , \label{e12 } \\ \dot{\tilde{\rho}}_{-1 } & = & -\left[\gamma_{- } + i(\omega_{21}-\lambda_{-}-\delta_{p})\right ] \tilde{\rho}_{-1 } -s\left(x_{4}\rho_{11}+x_{2}^{\ast}\rho_{--}\right)e^{-i\delta_{p}t } \nonumber \\ & & + i\omega_{p}\left[s\rho_{-+ } + c(\rho_{11}-\rho_{--})\right ] , \label{e13 } \\ \dot{\tilde{\rho}}_{1+}&= & -\left[\gamma_{+}^ { * } + i(\omega_{21}-\lambda_{+}+\delta_{p})\right]\tilde{\rho}_{1 + } -x_{2}e^{i\delta_{p}t}\rho_{-+ } \nonumber \\ & & + i\omega_{p}\left[s\left(\rho_{11 } - \rho_{++}\right ) + c\rho_{-+}\right ] , \label{e14}\\ \dot{\rho}_{-+ } & = & -(\gamma_{0}^{*}+i\omega_{r } ) \rho_{-+ } - x_{3}se^{-i\delta_{p}t}\tilde{\rho}_{1 + } + i\omega_{p}\left(s\tilde{\rho}_{-1 } + c\tilde{\rho}_{1+}\right ) , \label{e15}\end{aligned}\ ] ] where @xmath53 are the quantum interference terms , @xmath54 are the cavity modified damping rates between the dressed states , @xmath55 , \nonumber\\ \gamma_{- } & = & \gamma_{1}+\frac{g_{1}^{2}}{\kappa}\left(b_{3}^{\ast } + b_{4}^{\ast}\right ) + s^{2}\left[\gamma_{2}+\frac{g_{2}^{2}}{\kappa}\left(b_{0}+b_{1}\right)\right ] , \nonumber\\ \gamma_{+ } & = & \gamma_{1}+\frac{g_{1}^{2}}{\kappa}\left(b_{3}^{\ast } + b_{4}^{\ast}\right ) + c^{2}\left[\gamma_{2}+\frac{g_{2}^{2}}{\kappa}b_{2}\right]+s^{2}\frac{g_{2}^{2}}{\kappa}b_{0},\end{aligned}\ ] ] are the cavity modified damping rates of the coherence , with @xmath56 and @xmath57 are the dressed atom coherence in a rotating frame oscillating with frequency @xmath58 . equations ( [ e11])@xmath59[e15 ] ) are valid for any value of the cavity detuning @xmath60 , and the upper levels splitting comparable to half of the rabi frequency , i.e. for @xmath61 . physically , the approximation of @xmath61 corresponds to the case when the probe level @xmath62 is degenerate or nearly degenerate with respect to the dressed state @xmath63 . in this case , the resultant degeneracy gives rise to maximal quantum interference effects . moreover , under this approximation , @xmath64 and @xmath65 . the parameters appearing in the equations of motion have simple physical interpretations . the parameters @xmath66 are quantum interference terms . they contain contributions of both , spontaneously generated and cavity effects , which clearly illustrate an analogy between the cavity engineered and the spontaneously induced coherence @xcite . thus , the cavity with large decay rate strongly enhances quantum interference effects . the parameters @xmath67 represent the transition rates between the dressed states of the system and @xmath68 are the damping rates of the coherence . note that the parameters are dependent on the rabi frequency of the driving field and are resonant when the cavity frequency is tuned to @xmath69 . it means that spontaneous emission and quantum interference dominate at five frequencies . the sensitivity of the coefficients on @xmath60 is known in the literature as the purcell effect . thus , in the system considered here , both the purcell and the cavity - induced quantum interference effects play an important role in the dynamics and properties of the system . one can notice from ( [ e11])@xmath59([e15 ] ) that the coefficients in the differential equations are dependent on time . in fact , there is no reference frame in which the coefficients would be time independent . it is clear that the time dependence of the coefficients is brought here by the interference terms . as the result of the time dependence , special mathematical techniques must be employed to solve the set of the equations of motion . in the next section , we will solve the set of equations for the steady state density matrix elements using the floquet technique . our purpose of this paper is to demonstrate that the combined effect of the purcell and the cavity - induced quantum interference phenomena can create giant linear and nonlinear susceptibilities in the three - level system . note that the cavity - induced quantum interference effects are more flexible to the parameters than those induced by the spontaneously generated coherence . the latter depend solely on the angle between the dipole moments of the two atomic transitions . the former depend on the rabi frequency of the driving field , damping rates of the atomic transitions , and the detunings of the fields from their resonances . this makes the cavity system more practical for creation of quantum interference effects than that induced by spontaneously created coherence . it is well known that the response of the atomic medium to the probe field is governed by its polarization @xmath70 , which can be expressed in terms of the complex susceptibility @xmath71 or related to the elements of the density matrix of the system as @xmath72 where @xmath73 is the number density of the atoms , and @xmath74 is the atomic coherence on the probed transition . the task then is to determine the atomic coherence @xmath75 , or equivalently @xmath76 and @xmath77 , which could be found by solving the set of equations ( [ e11])@xmath59([e15 ] ) . in the stationary limit @xmath78 , we may set all of the derivatives to zero and obtain a set of algebraic equations for the density matrix elements . however , the set of equations retains the time dependence through the factors @xmath79 . therefore , to solve the system of equations ( [ e11])@xmath59([e15 ] ) , we employ the floquet method by expressing the density matrix elements as fourier series in terms of amplitudes that oscillate at the probe detuning and its harmonics . as we are interested in the response of the system to a weak probe field , we also make an expansion of the density matrix elements in terms of the powers of the probe field . these two decompositions combined together are given by the relation @xcite @xmath80 where the expansion in the powers of @xmath81 is given in terms of a dimensionless parameter @xmath82 that can take on values ranging continuously from zero ( no perturbation ) to one ( the full perturbation ) . since the atomic coherence on the probe transition oscillates as @xmath83 , the stationary properties of the first and third - order susceptibilities are determined by the harmonics @xmath84 and @xmath85 , respectively . therefore according to eq . ( 10 ) the susceptibilities @xmath86 and @xmath87 can be expressed in terms of the first and third order coherence of the probe transition as @xmath88 , \label{e22}\\ \chi^{(3 ) } & = & \frac{-2n_{a}|\vec{\mu}_{13}|^{4}}{3\hbar^{3}\varepsilon_{0}\omega_{p}^3 } \left[s(\rho_{1+})_{3}^{-1}-c(\rho_{1-})_{3}^{-1}\right ] .\label{e23}\end{aligned}\ ] ] the linear and nonlinear susceptibilities can be conveniently expressed in the form @xmath89 , \quad k=1,3 , \label{e24}\end{aligned}\ ] ] where we have introduced the normalized real and imaginary parts of @xmath90 that determine the index of refraction and the absorption coefficient , respectively . evidently , the normalized parts of @xmath90 are independent of the probe field strength @xmath91 . this allows the susceptibilities to be arbitrary large since the only approximation made here is an assumption of weak probe beam strengths . note that @xmath92 implies lossless propagation of the probe field , @xmath93 implies linear refraction of the probe beam , and @xmath94 implies nonlinear intensity dependent ( kerr ) refraction . upon substitution of ( [ e21 ] ) into ( [ e11])@xmath59([e15 ] ) and after comparing terms of the same powers in @xmath95 , we obtain an infinite set of equations for the fourier harmonics with time independent coefficients . despite of the complexity , the system of the coupled equations is easily solved for the steady state by an iteration in terms of the powers of the probe field amplitude . the analytical iterative solution for the @xmath96 order harmonics of the coherence appearing in ( [ e22 ] ) and ( [ e23 ] ) are of the form @xmath97}{(\gamma_{1}^{*}+in\delta_{p})(\gamma_{2}^{*}+in\delta_{p})-x_{2}x_{3}^ { * } } \nonumber \\ & & + i\omega_{p}\frac{x_{2}\left\{s[(\rho_{11})_{m-1}^{n+1}-(\rho_{++})_{m-1}^{n+1 } ] + c(\rho_{-+})_{m-1}^{n+1}\right\ } } { ( \gamma_{1}^{*}+in\delta_{p})(\gamma_{2}^{*}+in\delta_{p } ) -x_{2}x_{3}^ { * } } , \label{e25 } \\ ( \rho_{1-})_{m}^{n } & = & -i\omega_{p}\frac{c[(\rho_{11})_{m-1}^{n+1}-(\rho_{--})_{m-1}^{n+1 } ] + s(\rho_{+-})_{m-1}^{n+1}}{\gamma_{3}^{*}-in\delta_{p } } \nonumber \\ & & -\frac{s\left[x_{4}^{*}(\rho_{11})_{m}^{n } + x_{2}(\rho_{--})_{m}^{n}\right]}{\gamma_{3}^{*}-in\delta_{p } } , \label{e26}\end{aligned}\ ] ] where the analytical solutions for the auxiliary harmonics are quite lengthy and are listed in the appendix . it follows from the explicit solutions ( [ e25 ] ) and ( [ e26 ] ) that the magnitudes of the harmonics are proportional to @xmath98 , which ensures that their magnitudes are small even if the normalized susceptibilities @xmath99 and @xmath100 are large , since the probe beam strength is considered here to be weak , @xmath46 . this justifies the power expansion ( [ e21 ] ) . while ( [ e25 ] ) and ( [ e26 ] ) constitute an analytical solution to the susceptibility of the atomic medium , their form is algebraically complicated and there is little to be gained from a detailed dissection of these results . therefore , we will perform numerical analysis . we now proceed to perform detailed analysis of the the linear and nonlinear susceptibilities by graphically displaying the real and imaginary parts of @xmath86 and @xmath87 for a wide range of the important parameters . we are particularly interested in the possibility of creation of giant kerr nonlinearities accompanied by zero linear and nonlinear absorptions . in what follows , we assume for simplicity that the driving laser field is on resonance with the atomic transition @xmath101 , i.e. , @xmath102 . ( solid line ) , the linear @xmath103 ( dotted line ) and nonlinear @xmath104 ( dashed - dotted line ) absorption coefficients plotted as a function of the probe detuning @xmath105 for @xmath106 , @xmath107 , @xmath108 , @xmath109 , @xmath110 , and ( a ) @xmath111 , ( b ) @xmath112 , ( c ) @xmath113.,width=377 ] in fig . [ fig2 ] , we illustrate the variation of the real and imaginary parts of @xmath87 and the imaginary part of @xmath86 with the probe field detuning @xmath114 from resonance with the transition @xmath11 . we choose the rabi frequency of the driving field such that @xmath115 . in this particular case , the dressed state @xmath63 and the probe state @xmath62 are degenerated in the energies , which is the maximal quantum interference configuration . all of the parameters are measured in units of the damping rate @xmath116 through out these figures . part ( a ) of the figure shows the susceptibilities for @xmath111 . this corresponds to the cavity field tuned to the central component of the dressed transitions . we see that the susceptibilities exhibit resonance structures in the vicinity of the frequency @xmath117 . as we have already mentioned , at this frequency the quantum interference is maximal . the nonlinear ( kerr ) susceptibility is enhanced , but at the same time the linear and nonlinear absorptions are large . even at the frequency where the nonlinear absorption , im@xmath87 , vanishes , the linear absorption im@xmath86 is large with the magnitude comparable to the magnitude of the kerr nonlinearity . this is not desirable for a practical application since the probe beam could be completely absorbed over a short distant of propagation inside the atomic medium . therefore , we now proceed to check if one could achieve a large kerr nonlinearity accompanied by vanishing linear and nonlinear absorption by varying parameters of the system . a close inspection of the analytical expressions ( [ e25 ] ) and ( [ e26 ] ) shows that the absorption rate of the probe beam depends on the difference @xmath118 between the populations of the lower and upper levels of the probe transition which , on the other hand , depends on the detuning @xmath60 . thus , we expect that the transparency of the propagation of the probe beam could be improved by applying the purcell effect , i.e. by varying the detuning @xmath60 to match the cavity frequency with the frequency of one of the rabi sidebands of the driven transition . parts ( b ) and ( c ) of the figure show how the susceptibilities are modified when the cavity detuning @xmath60 is varied . there a few significant changes observed in the behavior of the susceptibility . firstly , the kerr nonlinearity becomes enhanced by few orders in magnitude when the detuning @xmath60 approaches the value @xmath113 , corresponding to the tuning of the cavity field to the rabi sideband of the driven transition . secondly , the kerr nonlinearity varies rapidly with the probe frequency . however , the most important change in the behavior of the susceptibility is that at the frequency , indicated by a dot d , where the kerr nonlinearity is maximal , the nonlinear absorption vanishes completely and the linear absorption is negligibly small . in other words , the system is transparent for the probe beam at the frequency where the kerr nonlinearity is maximal . we may conclude that by tuning the cavity field to one of the rabi sidebands , one can achieve a giant kerr nonlinearity accompanied by vanishing absorption . ( solid line ) , the linear @xmath103 ( dotted line ) and nonlinear @xmath104 ( dashed - dotted line ) absorption coefficients plotted as a function of the probe detuning @xmath105 for the same parameters as in fig . [ fig2](c ) but @xmath119 . the bottom part ( b ) presents the ratios @xmath120 ( solid lines ) and @xmath121 ( dotted lines ) plotted as a function of @xmath105 for the same parameters as in fig . [ fig2](c ) but two different values of @xmath122 : @xmath123 ( thick solid and dotted lines ) and @xmath119 ( thin solid and dotted lines).,width=377 ] we have seen that a combination of the maximal quantum interference and the purcell effect is crucial for creation of a giant kerr nonlinearity accompanied by vanishing absorption . to illustrate the importance of maintaining the maximal quantum interference , we now slightly detune the dressed state @xmath63 from the probed atomic state @xmath62 , so that the states become non - degenerate . it is well known , that quantum interference effects degrade when interfering energy states are non - degenerate . let us see how this can affect the kerr nonlinearity and the transparency of the atomic medium . in fig . [ fig3 ] , we plot the imaginary parts of @xmath86 and @xmath87 and the real part of @xmath87 for the same parameters as in fig . [ fig2](c ) , but with @xmath119 . in this case , the state @xmath63 is detuned from the state @xmath62 by @xmath124 , that is the cavity field is detuned from the dressed atom frequencies . it is easy to see from ( [ e18 ] ) that the effect of detuning the cavity field from the dressed atom frequencies is to reduce the magnitude of quantum interference terms . part ( a ) of the figure shows that the linear absorption is small at all frequencies , but within the region when the kerr nonlinearity is enhanced , the nonlinear absorption is very large . thus , the atomic medium becomes highly absorbing for the probe beam when the quantum interference effects are reduced . to illustrate further the effectiveness of the enhancement of the kerr nonlinearity by quantum interference , we plot in part ( b ) of the figure the ratios @xmath121 and @xmath120 for the presence and the absence of quantum interference . we see that at the frequency @xmath125 the ratios are maximal in the presence of quantum interference and vanish completely in the absence of quantum interference . note that the maxima of the ratios occur at frequencies slightly shifted from the resonance @xmath117 . this is because after adiabatically eliminating the cavity field operators in the bad cavity limit , the remaining cavity effects are not only to affect the atomic damping rates but also to induce a small energy shifts for the levels @xmath126 and @xmath1 . we may conclude that the enhanced kerr nonlinearity with relatively vanishing linear and nonlinear absorptions is a signature of the cavity - induced quantum interference effects . ( solid line ) , the linear @xmath103 ( dotted line ) and nonlinear @xmath104 ( dashed - dotted line ) absorption coefficients plotted as a function of the probe detuning @xmath105 for the same parameters as in fig . [ fig2](c ) but a very small damping rate on the probe transition @xmath127 . the bottom part ( b ) presents the kerr nonlinearity @xmath128 ( solid line ) , the linear @xmath103 ( dotted line ) and nonlinear @xmath104 ( dashed - dotted line ) absorption coefficients plotted as a function of the coupling constant @xmath129 for the same parameters as in fig . [ fig2](c ) but @xmath127 and @xmath130.,width=377 ] we now proceed to check the importance of other parameters of the system such as the atomic decay rate @xmath17 and the cavity field - atom coupling constant @xmath12 . figure [ fig4](a ) illustrates the susceptibility for the same parameters as in fig . [ fig2](c ) , but @xmath127 . it is evident that at this small damping rate , the linear absorption is zero at all frequencies , while the nonlinear absorption vanishes at two frequencies , indicated by dots e and f. at these frequencies the kerr nonlinearity is large . in particular , at the point f , the kerr nonlinearity is about one order higher in magnitude than that observed in fig . [ fig2](c ) for @xmath108 . evidently , the kerr nonlinearity can be enhanced and the linear and nonlinear absorptions kept zero by a proper choosing of the atomic decay rate on the probe transition . the dependence of the kerr nonlinearity and the absorption coefficients on the coupling constant @xmath12 is illustrated in fig . [ fig4](b ) . we show the imaginary parts of the susceptibilities @xmath86 and @xmath87 and the real part of @xmath87 as a function of @xmath12 for the probe detuning @xmath131 corresponding to the position of the maximum of @xmath128 . it is interesting to note that the linear absorption rate is zero independent of @xmath12 , while the nonlinear absorption varies from positive to negative and vanishes at @xmath132 . this shows that one can varies the magnitudes of the kerr nonlinearity and the absorption coefficients by a proper setting of the coupling constant . we close this section by a brief analysis of the adiabatic approximation and the range of the parameters used in our analytical treatment of the nonlinear dynamics of the system . one could object that the values of the parameters selected for plotting the figures are not in the range to fulfill the bad cavity limit of @xmath133 . in the first instance , we solve the master equation ( [ e1 ] ) numerically , using the quantum optics toolbox for matlab @xcite , for the steady - state values of the zeroth harmonics of the populations and coherence of the dressed states that determine the susceptibility of the system . we use the same values for the parameters as in fig . [ fig2](c ) , and the analytical and numerical results are listed in the table . it is evident that the discrepancies between the values of the density matrix elements obtained by the approximate solutions and corresponding exact numerical results are negligibly small . + [ cols="<,<,<",options="header " , ] + in the second , we plot in fig . [ fig5 ] the kerr nonlinearity together with the linear and nonlinear absorption coefficients for the same parameters as in fig . [ fig2](c ) but a significantly larger cavity damping rate , @xmath134 . we observe that the effects are qualitatively the same as those predicted for @xmath106 . the kerr nonlinearity attains maximal value at frequencies where the linear and nonlinear absorption are negligible . the only difference is in the numerical values of the magnitudes of the real and imaginary parts of the susceptibility . ( solid line ) , the linear @xmath103 ( dotted line ) and nonlinear @xmath104 ( dashed - dotted line ) absorption coefficients plotted as a function of the probe detuning @xmath105 for the same parameters as in fig . [ fig2](c ) but @xmath134.,width=377 ] we have studied the linear and nonlinear responses to a weak probe beam of a three - level atom coupled to a single - mode cavity and driven by a strong laser field . working in the bad cavity limit , we derived analytical expressions for the linear and nonlinear ( kerr ) susceptibilities . we have found that the joint effect of quantum interference and the purcell effect can lead to a giant kerr nonlinearity of the atomic medium accompanied by vanishing absorption . we have shown that the presence of maximal quantum interference is crucial for creation of the complete transparency of the atomic medium . the role of the significant parameters of the system has been discussed in details . we have shown that the creation of a giant kerr nonlinearity accompanied by vanishing absorption can be easily accomplished by a proper setting of the atomic decay rates or by a proper adjusting of the cavity field - atom coupling constants . the authors acknowledge financial support from the national natural science foundation of china ( under grants nos . 10674052 and 60878004 ) , the ministry of education under project ncet ( under grant no . ncet-06 - 0671 ) and srfdp(under grant no . 200805110002 ) , and the national fundamental research program of china ( under grant no . 2005cb724508 ) . 99 harris s e and hau l v 1999 phys lett . * 82 * 4611 see eg , special issue on quantum interference , 2002 j. mod . opt . * 49 * 1 schmidt h and imamolu a 1996 opt . lett . * 2 * 1936 tikhonenko v , christou j and luther - davies b 1996 phys . . lett . * 76 * 2698 harris s e and yamamoto y 1998 phys . lett . * 81 * 3611 nemoto k and munro w j 2004 phys . rev . lett . * 93 * 250502 ; jeong h 2005 phys . a * 72 * 034305 ; ottaviani c , _ et al . _ 2006 phys . a * 73 * 010301(r ) ; shapiro j h and razavi m 2007 new j. phys . * 9 * 16 imamolu a , _ et al . _ 1997 phys . 79 1467 ; werner m j and imamolu a 1999 phys . rev . a 61 011801(r ) niu y p , li r x , gong s q and liang x y 2005 opt . lett . * 30 * 3371 matsko a b , novikova i , welch g r and zubairy m s 2003 opt . lett . * 28 * 96 nakajima t 2000 opt . lett . * 25 * 847 niu y p and gong s q 2006 phys . a * 73 * 053811 yan x , wang l , yin b , jiang w , zheng h , song j and zhang y 2008 phys . a * 372 * 6456 zhu s y , chan r c f and lee c p 1995 phys . a * 52 * 710 ; zhu s y and scully m o 1996 phys . rev . * 76 * 388 ; agarwal g s 1997 phys . a * 55 * 2457 ; evers j and keitel c h 2002 phys . lett . * 89 * 163601 ; ficek z and swain s 2002 j. mod . opt . * 49 * 3 ; macovei m and keitel c h 2003 phys . lett . * 91 * 123601 ; li g x , evers j and keitel c h 2005 j. phys . b : at . mol phys . * 38 * 1435 lezama a , zhu y , morin s and mossberg t w 1989 phys . a * 39 * r2754 ; garraway b m and knight p l 1996 phys . rev . a * 54 * 3592 li g x 2000 opt . commun . * 184 * 267 zhou p 2001 phys . a * 63 * 023810 bermel p , rodriguez a , joannopoulos john d and soljai m 2007 phys . lett . * 99 * 053601 purcell e m 1946 phys . lett . * 69 * 681 ; kleppner d 1981 phys . * 47 * 233 ; macovei m , li g x , evers j and keitel c h 2007 phys . rev . lett . * 98 * 043602 ; bravo - abad j , _ et al . _ 2007 optics express * 15 * 16161 ; hamam r e , _ et al . _ 2008 optics express * 16 * 12523 brando f g s l , hartmann m j and plenio m b 2008 new j. phys . * 10 * 043010 ficek z and swain s , 2004 _ quantum interference and coherence : theory and experiments _ , springer , new york paspalakis e and knight p l 1998 phys . lett . * 81 * 293 zhu s y , chan r c f and lee c p 1995 phys . a * 52 * 710 fleischhaker r and evers j 2008 phys . a * 77 * 043805 tan s m http://www.qo.phy.auckland.ac.nz/qotoolbox.html in the appendix we present the analytical iterative solutions for the steady - state values of the fourier harmonics of the density matrix elements involved in the calculation of the linear and nonlinear susceptibilities . the @xmath96 harmonics are of the form @xmath135 + s(\rho_{-+})_{m-1}^{n-1}}{(\gamma_{3}+in\delta_{p } ) } \nonumber\\ & & + s\frac{x_{4}(\rho_{11})_{m}^{n } + x_{2}^{\ast}(\rho_{--})_{m}^{n}}{(\gamma_{3}+in\delta_{p } ) } , \nonumber \\ ( \rho_{-+})_{m}^{n}&= & i\omega_{p}\frac{x_{3}^{\ast}[s(\rho_{-1})_{m-1}^{n+1}+ c(\rho_{1+})_{m-1}^{n-1 } ] } { ( \gamma_{1}^{*}+in\delta_{p})(\gamma_{2}^{*}+in\delta_{p})-x_{2}x_{3}^ { * } } \nonumber\\ & & + i\omega_{p}\frac{(\gamma_{1}^{*}+in\delta_{p } ) \left\{s[(\rho_{++})_{m-1}^{n+1}-(\rho_{11})_{m-1}^{n+1 } ] -c(\rho_{-+})_{m-1}^{n+1}\right\}}{(\gamma_{1}^{*}+in\delta_{p})(\gamma_{2}^{*}+in\delta_{p})-x_{2}x_{3}^ { * } } , \nonumber\\ ( \rho_{+1})_{m}^{n } & = & i\omega_{p}\frac{(\gamma_{2}+in\delta_{p})\left[s(\rho_{1-})_{m-1}^{n-1 } + c(\rho_{+1})_{m-1}^{n+1}\right]}{(\gamma_{1}+in\delta_{p})(\gamma_{2}+in\delta_{p})-x_{2}^{*}x_{3 } } \nonumber\\ & & -i\omega_{p}\frac{x_{2}^{*}\left\{s[(\rho_{++})_{m-1}^{n-1}-(\rho_{11})_{m-1}^{n-1 } ] + c(\rho_{+-})_{m-1}^{n-1}\right\}}{(\gamma_{1}+in\delta_{p})(\gamma_{2}+in\delta_{p})-x_{2}^{*}x_{3 } } , \nonumber\\ ( \rho_{+-})_{m}^{n } & = & -i\omega_{p}\frac{x_{3}\left[s(\rho_{1-})_{m-1}^{n-1 } + c(\rho_{+1})_{m-1}^{n+1}\right ] } { ( \gamma_{1}+in\delta_{p})(\gamma_{2}+in\delta_{p } ) - x_{2}^{*}x_{3 } } \nonumber\\ & & + i\omega_{p}\frac{(\gamma_{1}+in\delta_{p})\left\{s[(\rho_{++})_{m-1}^{n-1}-(\rho_{11})_{m-1}^{n-1}]+c(\rho_{+-})_{m-1}^{n-1}\right\ } } { ( \gamma_{1}+in\delta_{p})(\gamma_{2}+in\delta_{p } ) - x_{2}^{*}x_{3 } } , \nonumber \\ ( \rho_{11})_{m}^{n } & = & \frac{-h_{4}u_{m-1}^{n\pm 1 } -h_{1}w_{m-1}^{n\pm 1}}{h_{1}h_{3}+h_{2}h_{4 } } , \nonumber \\ ( \rho_{--})_{m}^{n } & = & \frac{h_{2}u_{m-1}^{n\pm 1 } -h_{3}w_{m-1}^{n\pm 1}}{h_{1}h_{3}+h_{2}h_{4}},\end{aligned}\ ] ] where @xmath136 and @xmath137\nonumber \\ & & -i\omega_{p}\frac{x_{1}\left\{s(\rho_{-+})_{m-1}^{n-1 } -c\left[(\rho_{--})_{m-1}^{n-1}-(\rho_{11})_{m-1}^{n-1})\right]\right\}}{(\gamma_{3}+in\delta_{p } ) } \nonumber\\ & & + i\omega_{p}\frac{x_{1}^{*}\left\{s(\rho_{+-})_{m-1}^{n+1 } -c\left[(\rho_{--})_{m-1}^{n+1}-(\rho_{11})_{m-1}^{n+1})\right]\right\}}{(\gamma_{3}^{*}+in\delta_{p } ) } , \nonumber \\ w_{m-1}^{n\pm 1 } & = & -i\omega_{p}\left[s(\rho_{1+})_{m-1}^{n-1}-c(\rho_{1-})_{m-1}^{n-1 } -s(\rho_{+1})_{m-1}^{n+1 } + c(\rho_{-1})_{m-1}^{n+1}\right ] \nonumber\\ & & + i\omega_{p}\frac{x_{2}\left\{s(\rho_{-+})_{m-1}^{n-1}-c\left[(\rho_{--})_{m-1}^{n-1 } -(\rho_{11})_{m-1}^{n-1})\right]\right\}}{(\gamma_{3}+in\delta_{p})}\nonumber\\ & & -i\omega_{p}\frac{x_{2}^{*}\left\{s(\rho_{+-})_{m-1}^{n+1}-c\left[(\rho_{--})_{m-1}^{n+1 } -(\rho_{11})_{m-1}^{n+1})\right]\right\}}{(\gamma_{3}^{*}+in\delta_{p})}.\end{aligned}\ ] ]
we discuss a simple and experimentally realizable model for creation of enhanced kerr nonlinearities accompanied by vanishing absorption . the model involves a @xmath0-type atom subjected to a strong drive laser , a weak probe laser and coupled to a single - mode cavity field . working in the bad - cavity limit , we find that the simultaneous coupling of the cavity field to both atomic transitions creates a coherence between the transitions and thus can lead to quantum interference effects . we investigate the influences of the cavity field frequency , the cavity field - atom coupling constants and the atomic decay constants on the linear and the third - order ( kerr ) nonlinear susceptibilities . we predict giant kerr nonlinearities with vanishing absorption and attribute this effect to the combination of the purcell effect and the cavity - induced quantum interference .
this paper is part of a global analysis of star counts developed to constrain consistently scenarii of galaxy formation and evolution.the central tool of this approach is the `` besanon '' model of population synthesis . this model is gradually tuned to fit an increasing number of observational constraints while keeping compatibility with previous fits and theoretical prescriptions . in the present paper we address the problem of the halo ( dark or visible ) by trying to compare the properties of the spheroid population ( the visible halo ) with the dark matter halo , as traced by microlensing at high galactic latitudes and by the rotation curve . if the dark matter is made at least partly of stellar remnants , as shown by recent statistics of microlensing at high galactic latitudes @xcite , the density trend of this matter should be close to a power law with index of 2 ( as expected from a flat rotation curve ) . it is natural to think of a similar shape for the stellar spheroid . constraints on the overall shape of the dark halo are poor . cosmological simulations of halo formation generally predict that halos are flattened by about c / a @xmath5 0.7 @xcite . but the axis ratio depends on how much the halo matter is dissipative , the more dissipative , the flatter the halo . direct determinations of the dark matter distribution in polar ring galaxies show flattened halos with c / a @xmath5 0.5 @xcite . concerning the spheroid population , most previous analyses suggest rather steep density slopes with power indices between 3.0 and 3.5 . however , these analyses are based on rather small samples of well identified tracers . the estimated flattening also cover a wide range between 0.6 to 1.0 : the distribution of galactic globular clusters appears to be well fitted by a power law density with index @xmath6 and flattening of 1 . hawkins rr lyrae observations @xcite showed @xmath7 with a flattening of 0.9 . saha @xcite , using a spherically symmetric model , found @xmath8 out to 25 kpc but then the rr lyrae density falls off more rapidly beyond 25 kpc . another study of rr lyrae by wetterer @xcite showed that a spherically symmetric model yields @xmath9 whereas an ellipsoidal distribution yields @xmath6 . sluis @xcite counted blue horizontal branch ( bhb ) stars and rr lyrae and found @xmath10 and @xmath11 . still from bhb star counts , sommer - larsen @xcite derived @xmath12 and @xmath8 up to 40 kpc , preston @xcite found that @xmath13 increases from 0.5 to 1 up to 20 kpc with @xmath14 . soubiran @xcite showed that @xmath15 is compatible with the kinematical behavior of a star sample near the north galactic pole . k dwarf counts with hst yield @xmath16 and @xmath17 @xcite . all of these studies were based on a few hundred objects at most . in order to find new constraints on the spheroid density law , we undertook a photometric and astrometric sample survey in various galactic directions . we complemented these data with existing deep photometric star counts in several high and intermediate latitude fields . most such counts contain large numbers of halo dwarfs , but they can not be distinguished from thick disc dwarfs by their colours but at faint magnitudes . since no large optical surveys were available at magnitude fainter than 20 , we used heterogeneous data coming from various studies ( often of extragalactic aim ) in various photometric systems . the population synthesis model used here permits to perform a global analysis of these heterogeneous data , since observational data can be simulated in each field with the true observational conditions ( photometric system , errors and selection effects ) . the synthetic approach allows also to estimate the biases and expected contaminations by other populations . in section 2 we describe the model of population synthesis and external constraints on the spheroid population . in section 3 we describe the data sets and the comparison method . in section 4 we discuss the results and their implications for the dark matter halo . we have used a revised version of the besanon model of population synthesis . previous versions were described in bienaym et al . and haywood et al . . the model is based on a semi - empirical approach , where physical constraints and current knowledge of the formation and evolution scenario of the galaxy are used as a first approximation for the population synthesis . the model involves 4 populations ( disc , thick disc , halo and bulge ) each deserving a specific treatment . the bulge population which is irrelevant for this spheroid analysis will be described elsewhere . a standard evolution model is used to produce the disc population , based on a set of usual parameters : an initial mass function ( imf ) , a star formation rate ( sfr ) , a set of evolutionary tracks ( see haywood et al . , 1997 and references therein ) . the disc population is assumed to evolve during 10 gyr . a set of imf slopes and sfr s are tentatively assumed and tested against star counts . the tuning of disc parameters against relevant observational data was described in haywood et al . . the model fixes the distribution of stars in the space of intrinsic parameters : effective temperature , gravity , absolute magnitude , mass and age . these parameters are converted into colours in various systems through stellar atmosphere models corrected to fit empirical data . while some errors still remain in the resulting colours for some spectral types , the overall agreement is good in the major part of the hr diagram . since the haywood et al . model was based on evolutionary tracks at solar metallicities , inverse blanketing corrections are introduced to give to the disc a metallicity distribution in agreement with twarog @xcite age / metallicity distribution ( mean and dispersion about the mean ) . the model returns the present - day distribution of stars as a function of intrinsic parameters in a unit volume column centered at the sun position . since the evolution model does not account for orbital evolution , stars are redistributed in the reference volume over the z axis . the key for redistributing stars along the z - axis is age : an empirical relation associates z velocity dispersions to ages . then the boltzmann equation is used to convert z velocity distributions into z density . the model is dynamically self - consistent in the sense that the potential used in the boltzmann equation is the one generated by the total mass distribution of stellar populations . the self consistency is established iteratively . we slice the disc populations into seven isothermal populations of different ages , from 0 to 10 gyr . each sub - population ( except the youngest one , which can not be considered as relaxed ) has its velocity dispersion imposed by the age / velocity dispersion relation . we then deduce the scale height of each sub - population using the boltzmann equation . the overall scheme is described in bienaym . resulting density laws are used to correct the evolution model distribution in and off the plane , then to compute the stellar densities all over the galaxy . a detailed analysis of the thick disc population from photometric and astrometric star counts has been given elsewhere . the kinematics , metallicity , and density law were measured allowing us to constrain the origin for this population . in this series of papers , evidence was given that the majority of thick disc stars should originate from a merging event at the beginning of the life of the thin disc , after the first collapse . one or several satellite galaxies have heated the thin disc , then the gas re - collapsed and reformed a new thin disc . in the population synthesis process , the thick disc population is modeled as originating from a single epoch of star formation . we use bergbush & vandenberg @xcite oxygen enhanced evolutionary tracks . no strong constraint exists on the thick disc age until now . we assume an age of 11 gyr , which is slightly older than the disc and younger than the halo . the initial mass function is modeled by a simple power law with a slope about @xmath18 , referring to the notation @xmath19 . the thick disc metallicity can be chosen between and dex in the simulations . the standard value of dex is usually adopted , following in situ spectroscopic determination from gilmore et al . @xcite and photometric star count determinations . the low metallicity tail of the thick disc seems to represent a weak contribution to general star counts @xcite . it was neglected here . an internal metallicity dispersion among the thick disc population is allowed . the standard value for this dispersion is 0.25 dex . no evidence has been found for a significant metallicity gradient in the thick disc population ( robin et al , 1996 ) . the thick disc density law is assumed to be a truncated exponential : at large distances the law is exponential . at short distances it is a parabola . this formula ensures the continuity and derivability of the density law ( contrarily to a true exponential ) and eases the computation of the potential . the scale height of the exponential can vary between 600 and 2600 pc . the standard value , 760 pc , has been obtained from star count fitting in various directions ( robin et al , 1996 ) . nevertheless , it can be shown that star counts when restricted to a small number of galactic directions and a small magnitude range do not give a strong constraint on the scale height , but rather on the parameter : ( local density)@xmath20(scale height)@xmath21 . at present there is no accurate determination of the thick disc density in the solar neighbourhood , independently from the scale height . but reasonable values range between 700 to 1200 pc for the scale height and 1 to 4% for the local density relative to the thin disc . we assume a homogeneous population of spheroid stars with a short period of star formation . we thus use the bergbush and vandenberg @xcite oxygen enhanced models , assuming an age of roughly 14 gyr ( until more constraints on the age are available ) , a mean metallicity of -1.7 dex and a dispersion of 0.25 about this value . no galactocentric gradient is assumed . the imf has to be constrained either from globular clusters ( if they are representative of the spheroid population ) or from deep star counts . this point is discussed in section 3 . the density of spheroid stars is modeled by a power law : @xmath22 where @xmath23 is the local density , @xmath24 is the power law index and @xmath25 is the flattening . the local density can be constrained by local measurements of high velocity stars , or by remote counts of giants ( spectroscopically selected ) or dwarfs ( photometrically selected ) . the local density can not be determined independently from the other density parameters with our limited number of data sets . thus we have used independent constraints from the literature on the local spheroid density . the local stellar spheroid density , @xmath23 , is bounded by observational data on halo dwarfs and giants . figure [ fig - flum ] shows the luminosity function obtained by different authors . we only selected recent results obtained in good conditions from sufficiently large samples . bahcall et al @xcite and gizis & reid @xcite derived their values from high proper motion dwarf samples . dahn et al @xcite determined accurate parallaxes for local late - type subdwarfs and deduced the local luminosity function of halo stars in the absolute visual magnitude range 9 to 14 . these three results are biased by the kinematic selection . they took the bias correction into account but this correction is model dependent and introduces an unknown uncertainty into the result . we expect that the differences between the three measurements rely upon this correction . on the giant side , morrison @xcite used a non kinematically - biased sample of halo giants , selected from their metallicity to estimate the spheroid local density . in figure [ fig - flum]a we show the luminosity function from bergbush and vandenberg ( 1992 ) for a population of 14 gyr with a metallicity of -1.75 and an imf slope @xmath0 of 2 . if we let the local halo density vary from a factor of 0.75 to 1.25 relative to this reference model ( dotted lines in figure [ fig - flum]a ) , we get a good agreement with the specified observations given their uncertainties . in the next section we allow the local density to vary within these limits . obtaining good constraints on the spheroid density law requires a good photometric accuracy . this generally depends on using ccd detectors on large telescopes , on fields as wide as possible to cover large samples , and a large range of galactocentric distances . this can be obtained with a number of data sets at various galactic latitudes and longitudes . we have collected such data sets from the literature . most have been made for extragalactic purposes . the main data characteristics are summarized in table 1 . the photometric systems are close to the johnson - cousins system . spheroid star selection was based on their magnitude and colour ( either b - v , v - r , or v - i , depending on the available observations ) , in order to avoid presence of contamination by other populations . aiming at model independent results , the model was used essentially to select colour and magnitude ranges and fields where the contamination by thick disc stars remains negligible under any reasonable thick disc hypothesis . for this reason all data brighter than magnitude 20 at intermediate and low latitudes were excluded . a small number of disc white dwarfs is also present in the selection but the proportion is at most a few percent and has no consequence on the result . our survey program include at the moment two fields , one towards the north galactic pole , another at intermediate latitude ( l=150,b=60 ) . the ngp field is the deepest up to now : it is complete and free from galaxy contamination up to magnitude 24 . a full description of these data sets will be given in a forthcoming paper . the other selected data sets are the six fields of the dms survey @xcite observed in v and r bands at medium latitude , 4 fields from the canada - france redshift survey ( cfrs , @xcite ) dedicated to galaxy counts , two fields from the koo and kron investigation for quasars @xcite , another field from reid and majewski near the north galactic pole @xcite . the absolute visual magnitude of halo stars in the selected samples ranges between 3 and 8 , except our north galactic pole field which reaches m@xmath26 . all these fields taken together cover a large part of the ( r , z ) plane , as can be seen in figure [ fig - rz ] where the distributions in r and z of 90% of halo stars in each field of view are drawn . [ cols= " < , < , < , < , < , < " , ] spheroid models with a local density 0.75 @xmath23 and 1.0 @xmath23 give the maximum likelihood , well in agreement with the previous result . however the model with 0.50 @xmath23 is noticeably worse . eventually , the resulting imf slopes do not depend significantly on the assumed density laws and the likelihood is well peaked around the maximum indicating a robust determination . we conclude that the imf of the halo , in the mass range [ 0.1 , 0.8 ] is : @xmath27 while the imf slope of the thick disc seems to be slightly smaller and similar to the disc s @xcite . these values do not account for binarity . thus the true imf should slightly steepen . we leave the value uncorrected until more data are available on the binary fraction among low mass spheroid stars . this result is the first direct measurement of the mass function of field star spheroid with a good statistics , thanks to the wide field of the ccd mosaic . several previous determinations used kinematically selected samples ( see section 2.4 for references ) or deep fields . but the latter were limited to narrow fields : the first attempt by richer & fahlman @xcite lead to a very steep imf slope of @xmath28 which had given hope for a dark matter halo of brown dwarfs . later results have given shallower slopes but the uncertainties were not significantly decreased . gould et al . @xcite analyzed a sample of 166 stars in 53 hst wfpc fields , making difficult the de - correlation between structural parameters of the spheroid and its mass function . they found a luminosity function down by a factor two from the present one and deduced an imf slope of @xmath29 ( in our notation ) . their result relies upon the assumption that the spheroid has a mean metallicity of -1.0 , which looks too high considering most direct measurement of its abundances . this high metallicity induces an overestimate of the luminosity at a given colour , hence of the distance , as well as an overestimate of the mass relatively to a smaller assumed metallicity . if we independently check the results obtained in inner fields and in outer fields , we are able to search for solutions with varying power law index and flattening over the galactic radius . contrarily to preston et al . ( 1991 ) we find no evidence for varying power law index or flattening . however , a round spheroid is ruled out by the inner field data as well as by fields at low latitudes . thus our results are compatible with a true power law and a constant flattening all along the tested galactic radius . when comparing data sets from different sources in close galactic fields , discrepancies appear which are larger than expected on the basis of pure random noise . this may be due either to data incompleteness , or to systematic errors in the photometry ( including mismatch of the standard photometric system ) , or to true inhomogeneities in the spheroid distribution . currently available data are not sufficient to discriminate between these different causes . homogeneous wide field surveys will be necessary to clarify these aspects . the scope of the current investigation is for this reason limited to large scale average characteristics . the blue peak at these magnitudes may be contaminated by disc white dwarfs or by thick disc main sequence stars . the former are very few compared to the density of the halo . the contamination by disc white dwarfs , as determined by the model , is at most 5% in the magnitude range 22 - 24 . the contamination from the thick disc has been estimated using our best fit thick disc model as adjusted on medium deep star counts . the contamination can reach about 30% in the magnitude range 18 - 20 but becomes negligible at magnitude larger than 20 as seen in figure [ fig - contamin - mag ] . hence , we do not take into account magnitudes lower than 20 in our study . had the thick disc contamination been underestimated , then the contribution assigned to the halo in the blue peak would be too large , resulting in a possible distortion of the density law . in order to evaluate how this would affect our conclusions , we have investigated different thick disc models which could fake the halo contribution to the blue peak . attempts were limited to realistic thick discs roughly fitting the red peak . we have selected two extreme thick disc parameters for which the contamination to the blue peak becomes significant . a thick disc with a local density of 3.9% , a scale height of 1150 pc and an imf slope of 1 . ( referred to as model b ) . a thick disc with a scale height of 2 kpc , a local density of 0.5% and an imf slope of 1.75 . ( model c ) . with such thick disc models , the process of adjusting the spheroid density law parameters end up to tiny local density of about 25% of the standard value and very small power law index of the order of 1.5 . surprisingly , the fit is good on star counts up to magnitude 22 , showing that a large range of parameters can reproduce a wide set of star counts . however at magnitude 22 - 24 , model b and c are unable to reproduce the counts , as exemplified in figure [ fig - contaminbc ] , where at the top star counts at the pole in the magnitude range 22 - 24 are overestimated in the blue peak , and in the range 18 - 22 ( bottom ) the fit is still acceptable . so , the degeneracy between thick disc models and halo parameters holds only if star counts are not deep enough . keeping reasonable values for the thick disc parameters leads to a small contamination with no risk of underestimation of the halo density . if the thick disc contribution is higher than expected from standard models then we would overestimate the local halo density and power law index , strengthening our conclusion towards a flat spheroid with a small power law index . we have shown that , with the data available up to now , the spheroid star distribution follows a power law with an index smaller than previously thought . it is moderately flattened , as already found by several investigations . the best fit power law index is found 2.44 for a flattening of 0.76 . we can not exclude a spheroid population having a power law as low as 2 and an axis ratio of 0.5 at the 2 sigma level . assuming a high exponent of 3.5 , as suggested by some globular cluster and rr lyrae data , model predictions deviate significantly from observations in the external parts of the galaxy whatever reasonable flattening is adopted . the imf slope of the spheroid is found to be @xmath30 , value which gives a local density of 1.64 10@xmath2 stars pc@xmath3 and a mass density of 4.15 10@xmath4 pc@xmath3 for the stellar halo , yet this value ignores possible old white dwarfs . with this slope the expected mass density of brown dwarfs in the halo makes a negligible part of the dark matter halo , as already estimated from microlensing surveys . recent searches for ancient halo white dwarfs have given a hope to identify the microlensing events with such objects @xcite . ibata et al . @xcite conclude that old white dwarfs may constitute a significant fraction of about 10% of the dark matter halo . in the mean time , microlensing experiments have narrowed the range of the estimated halo baryonic fraction to 20 to 50% @xcite . these two results are well in agreement according to the uncertainties . so , as star count data progresses in depth and extent , the picture of the spheroid star population that comes out points to characteristics quite compatible with what we know about the distribution of baryonic dark matter if it is made of stellar remnants , suggesting a common dynamical origin . the visible spheroid and its heavy counterpart of dark remnants can make a significant but not dominant part of the so - called dark matter halo . alcock , c. , allsman , r. , alves , d. , axelrod , t. , becker , a. , bennett , d. , cook , k. , dalal , n. , drake , a. , freeman , k. , geha , m. , griest , k. , lehner , m. , marshall , s. , minniti , d. , nelson , c. , peterson , b. , popowski , p. , pratt , m. , quinn , p. , stubbs , c. , sutherland , w. , tomaney , a. , vandehei , t. , and welch , d. : 2000 , , c. , allsman , r. a. , alves , d. , ansari , r. , aubourg , e. , axelrod , t. s. , bareyre , p. , beaulieu , j. p. , becker , a. c. , bennett , d. p. , brehin , s. , cavalier , f. , char , s. , cook , k. h. , ferlet , r. , fernandez , j. , freeman , k. c. , griest , k. , grison , p. , gros , m. , gry , c. , guibert , j. , lachieze - rey , m. , laurent , b. , lehner , m. j. , lesquoy , e. , magneville , c. , marshall , s. l. , maurice , e. , milsztajn , a. , minniti , d. , moniez , m. , moreau , o. , moscoso , l. , palanque - delabrouille , n. , peterson , b. a. , pratt , m. r. , prevot , l. , queinnec , f. , quinn , p. j. , renault , c. , rich , j. , spiro , m. , stubbs , c. w. , sutherland , w. , tomaney , a. , vandehei , t. , vidal - madjar , a. , vigroux , l. , and zylberajch , s. : 1998 , , l9 , c. , allsman , r. a. , alves , d. , axelrod , t. s. , becker , a. c. , bennett , d. p. , cook , k. h. , freeman , k. c. , griest , k. , guern , j. , lehner , m. j. , marshall , s. l. , peterson , b. a. , pratt , m. r. , quinn , p. j. , rodgers , a. w. , stubbs , c. w. , sutherland , w. , welch , d. l. , and the macho collaboration : 1997 , , 697 , e. , bareyre , p. , brehin , s. , gros , m. , lachieze - rey , m. , laurent , b. , lesquoy , e. , magneville , c. , milsztajn , a. , moscoso , l. , queinnec , f. , rich , j. , spiro , m. , vigroux , l. , zylberajch , s. , ansari , r. , cavalier , f. , moniez , m. , beaulieu , j. p. , ferlet , r. , grison , p. , madjar , a. v. , guibert , j. , moreau , o. , tajahmady , f. , maurice , e. , prevot , l. , and gry , c. : 1993 , , 623 dahn , c. c. , liebert , j. , harris , h. c. , and guetter , h. h. : 1995 , in c. g. tinney ( ed . ) , _ the bottom of the main sequence and beyond _ , pp 239248 , springer , proceedings of the eso workshop held in garching , germany , 10 - 12 august 1994
as part of a global analysis of deep star counts to constrain scenarii of galaxy formation and evolution , we investigate possible links between the galactic spheroid and the dark matter halo . a wide set of deep star counts at high and intermediate galactic latitudes is used to determine the large scale density law of the spheroid . assuming a power density law , the exponent , flattening , local density and imf slope of this population are estimated . the estimation is checked for robustness against contamination of star counts by the thick disc population . contamination effects are derived from a model of population synthesis under a broad variety of thick disc parameters . the parameter fit is based on a maximum likelihood criterion . the best fit spheroid density law has a flattening of 0.76 , a power index of 2.44 . there is a significant degeneracy between these two parameters . the data are also compatible with a slightly less flattened spheroid ( c / a = 0.85 ) , in combination with a larger power index ( 2.75 ) . a flatter spheroid ( c / a = 0.6 ) with a power index of 2 is not excluded either . we also constrain the spheroid imf slope @xmath0 to be 1.9 @xmath1 0.2 , leading to a local density of 1.64 10@xmath2 stars pc@xmath3 and a mass density of 4.15 10@xmath4 pc@xmath3 . with this slope the expected mass density of brown dwarfs in the halo makes a negligible part of the dark matter halo , as already estimated from microlensing surveys . so , as star count data progresses in depth and extent , the picture of the spheroid star population that comes out points to a shape quite compatible with what we know about the distribution of baryonic dark matter if it is made of stellar remnants , suggesting a common dynamical origin .
indirect dark matter detection is the search for the annihilation or decay products of particle dark matter instead of _ directly _ detecting the dark matter , _ indirect _ detection aims at pinpointing the non - standard origin of a cosmic - ray or a gamma - ray signal based upon the messenger s spectral or angular features ( see the relevant chapters of ref . @xcite for a collection of recent reviews ) . examples of indirect detection with gamma rays are searches for a peculiar gamma - ray signal from nearby dwarf spheroidal galaxies ( see e.g. @xcite ) , from nearby clusters ( e.g. @xcite ) , from the center of the galaxy ( e.g. @xcite ) , in the extragalactic gamma - ray background ( e.g. @xcite ) or in the diffuse background angular distribution ( e.g. @xcite ) ; high - energy neutrinos from the center or of the sun or of the earth ( see e.g. @xcite ) , the search for features in the spectrum of positrons or antiprotons ( e.g. @xcite ) , and the search for antidetuerons ( e.g. @xcite ) are further examples of indirect dark matter detection channels . this is an especially exciting point in time for the field of indirect dark matter detection . the fermi large area telescope ( lat ) is well into its third , successful year of science operations . crucial aspects of the gamma - ray sky for indirect dark matter searches , such as the issue of the galactic and extra - galactic diffuse background , are undergoing close scrutiny with significant progress well under - way @xcite . research and development is also in progress for the next generation gamma - ray detector , a ground - based cherenkov telescope array ( cta ) , whose significantly lower low - energy threshold will enormously benefit its use for dark matter searches . the icecube neutrino telescope has recently been completed , with a thickly - instrumented portion ( deepcore ) that makes it an even better dark matter detection instrument than originally designed @xcite . finally , the ams-02 payload has recently been successfully deployed on - board the international space station , and science results will soon be delivered . the weakly - interacting massive particle ( wimp ) paradigm for the nature of dark matter is a highly compelling scenario , as it relies on generic weak - scale physics and on the well - known thermal freeze - out mechanism for the production of a relic with just the right density to match the inferred dark matter abundance in the universe . however , the so - called wimp miracle does not provide an explanation to why the universal density of baryons @xmath0 ( in units of the critical density , @xmath1 ) is tantalizingly close to the density of dark matter @xmath2 , so much as to pose the veritable `` coincidence problem '' of explaining why @xmath3 . a general class of models posits that the dark matter relic density is in fact set by an asymmetry between dark matter and anti - dark matter particles , in full similarity to the baryonic relic density ( e.g. @xcite ) . if the two asymmetries are related to each other ( for example if there exists a `` transfer mechanism '' between the dark matter and the baryonic sector ) , then generically @xmath4 gev , with @xmath5 the proton mass and @xmath6 a model - dependent number of order unity . these models have been dubbed `` asymmetric dark matter '' models in ref . @xcite , and received a considerable amount of attention recently , in part because of direct detection results that might point to a light ( with a mass of a few gev ) dark matter candidate . while asymmetric dark matter models give rise to an interesting phenomenology ( see e.g. @xcite ) , including new contributions to proton decay , and , as mentioned , can even explain anomalies in direct dark matter experiments ( e.g. @xcite ) , by definition , asymmetric dark matter does not annihilate ( or does so at a very suppressed rate , should there be a left - over symmetric component ) . one might thus naturally wonder whether indirect dark matter detection is or not entirely doomed in asymmetric dark matter scenarios ( see however @xcite ) . in this paper we consider a process that was , in an entirely different astrophysical setting , suggested some time ago by bloom and wells @xcite . there , the authors suggested that high - energy electrons in the jets of active galactic nuclei ( agn ) could up - scatter dark matter particles , possibly radiating a photon in the final state . estimating , with the techniques of effective field theory , the relevant cross section , and estimating the flux of electrons and the density of dark matter particles , the authors concluded that ( 1 ) the level of the signal was several orders of magnitude below the sensitivity of the lat , and ( 2 ) no specific `` smoking - gun '' spectral features were present in the signal . in ref . @xcite we recently studied the same process originally proposed in @xcite , improving on the previous analysis in a number of ways . first , we carried out the complete elementary process calculation , including resonance and threshold effects , for two particle dark matter models ( supersymmetric neutralinos and the lightest kaluza - klein of universal extra dimensions ) for both cosmic - ray electrons and protons in the agn jet ; second , we re - evaluated the agn jet components and the dark matter density in view of recent observational and theoretical - modeling results . we concluded that the process can be resonant , and that the resonance entails on the one hand a larger final - state photon flux , and on the other hand a distinctive spectral feature , corresponding to a certain , well - understood kinematic condition . as a result , we predicted that for some dark matter models the process of dark matter scattering off of agn jet particles can be observable by the fermi lat . here , we observe that our own galaxy is populated by cosmic rays ( hadrons and heavy nuclei , electrons , positrons and neutrinos ) that might interact with the galactic dark matter a process that might be `` visible '' with a gamma - ray telescope such as fermi lat if it radiates a photon in the final state . evidently , the elementary processes we considered in detail in ref . @xcite are precisely the same for the case of scattering off of galactic cosmic rays ( albeit a few technical differences will be of importance , as we point out later ) . a further possibility , that we leave for future work , is that dark matter - cosmic ray scattering lead to inelastic processes producing e.g. neutral pions , eventually resulting in gamma rays . note that inelastic interactions between cosmic - ray protons and dark matter particles have been also recently considered in ref . @xcite and @xcite as a possible explanation to the origin of the observed `` knee '' in the spectrum of galactic cosmic rays , for dark matter models in the context of extra dimensions with a low fundamental scale of gravity . limits on the interaction cross section of dark matter off of hadronic cosmic rays have also been considered in ref . @xcite in the context of models of strongly interacting dark matter ; the strongest constraints were found to arise from effects on the predicted abundances of light elements in big bang nucleosynthesis . in this study we first consider , in sec . [ rough ] , the process of dark matter scattering off of cosmic rays in the galaxy for a generic wimp model , and we compare the expected luminosity from final - state radiated photons with what expected for wimp pair - annihilation ; we also consider the same process for two additional dark matter models that received some attention recently : dipolar dark matter and `` luminous '' dark matter . we then present the core of our numerical results in sec . [ core ] , and separately give details for the case of cosmic - ray electrons ( [ ele ] ) and protons ( [ pro ] ) . we present a discussion and our conclusions in the final sec . [ sec : disc ] . in this section we estimate , for generic wimp models as well as for dipolar and luminous dark matter models , the expected luminosity for the process where a photon is emitted in the final state of a cosmic - ray scattering off of dark matter . we start by comparing , for the case of ( symmetric ) wimp dark matter models where pair - annihilation can occur , the luminosity of the final - state photon emission from cosmic ray scattering off of dark matter to the well - known dark matter pair - annihilation rate . both processes can be thought of as occurring on the same target particles : galactic dark matter . we thus construct an order - of - magnitude estimate for the relevant rates , per unit target . these are defined here as follows : * for dark matter annihilation , @xmath7 with @xmath8 the pair annihilation cross section , @xmath9 the number density , and @xmath10 the relative dark matter particle velocity in the galaxy * for cosmic ray - dark matter scattering , @xmath11 with @xmath12 the locally measured differential cosmic ray electron flux , averaged over angles , and @xmath13 the relevant scattering cross section . we estimate a weak - scale pair - annihilation cross section @xmath14 and we employ a local dark matter density @xmath15 and a dark matter mass of 100 gev to estimate @xmath16 . finally , we take @xmath17 . with these numbers , we find @xmath18 for the case of cosmic ray - dark matter scattering we employ @xmath19 as measured and reported by the lat collaboration @xcite ( we consider cosmic - ray electron energies around 1 gev given that we are interested in final - state gamma rays with energies in the fermi lat range , thus above 0.1 gev , and because at larger energies the cosmic - ray spectrum declines very steeply ; we thus expect the energy range that contributes the most to the signal we are interested in to be in the gev range ) , and we estimate @xmath20 with possible resonances enhancing these numbers by 2 - 4 orders of magnitude , according to the findings of @xcite ( see e.g. fig . 3 ) . as a result , @xmath21 this calculation only employs the local dark matter density and cosmic - ray electron flux : both these quantities are larger the closer one gets to the center of the galaxy . this order of magnitude estimate , however , implies that in the generic case where the dark matter pair - annihilates and there are no symmetries or kinematic effects suppressing the natural pair - annihilation rate for a weak - scale weakly - interacting particle , the effect under investigation here is suppressed by 3 to 7 orders of magnitude compared to pair - annihilation . notice also that the number of photons produced in each _ annihilation event _ is typically on the order of 10 - 100 , as opposed to the single photon that scattering off of cosmic ray would produce _ per event _ ( an exception to this is given by inelastic processes leading to neutral pions in the final state ) . since the detectable quantity here is not the event - rate , but the actual gamma rays produced in the interactions , the suppression in the predicted gamma - ray flux is thus generically larger by a further 1 to 2 orders of magnitude . if , however , effects such as co - annihilation or velocity dependence of the pair - annihilation are in place , the two processes could produce a comparable flux of gamma rays . in addition , if dark matter can not pair - annihilate because of discrete symmetries or because of lack of annihilation partners , such as in asymmetric dark matter models @xcite , then scattering off of cosmic ray electrons with a final - state photon emission might be one , if not the only way to detect a signal from dark matter with gamma - ray telescopes . it is interesting to investigate two more dark matter scenarios that have received some attention recently : dipolar dark matter ( see e.g. @xcite ) , and luminous dark matter @xcite . the former arises as dark baryons in hidden - sector gauge mediation supersymmetry breaking ( see e.g. @xcite ) , or in technicolor models ( see e.g. @xcite ) . while the relic density of this class of dark matter candidates is produced non - thermally , often a mechanism relates the dark baryon to the standard - model baryon density , thereby providing an explanation to the coincidence problem @xmath22 @xcite . luminous dark matter , on the other hand , is an interesting proposal in light of recent direct detection results @xcite . we estimate the dark matter scattering off of cosmic ray electrons for the case of dipolar dark matter assuming that the dark matter is a fermion @xmath23 of mass @xmath24 gev , with a dipolar interaction cross section of the type @xmath25 we estimate the cross section off of cosmic - ray electrons with a typical energy scale around 1 gev , and consider for simplicity the internal bremsstrahlung approximation of eq . ( 17 ) of ref . @xcite for the final state photon radiation , with an outgoing photon energy on the gev range . notice that this process is never resonant . we obtain the estimate @xmath26 this entails that unless the dark sector scale @xmath27 is very low and the dark matter is very light the dipolar dark matter case is less promising than the wimp case . in the context of luminous dark matter @xcite , the dipolar operator connects two majorana fermions , @xmath28 ( the dark matter particle ) and @xmath29 , with @xmath30 gev and @xmath31 and @xmath32 kev chosen to provide an explanation to the observed dama / libra spectrum . the relevant interaction term is @xmath33 here the picture is one where scattering of @xmath28 off of cosmic ray via the dipolar interaction operator produces meta - stable @xmath29 particles , subsequently decaying into @xmath34 with the photon energy in the final state @xmath35 kev in the x - ray regime . we estimate the cross section for the process @xmath36 for @xmath37 gev to be on the order of @xmath38 this cross section would produce a rate of 3.3 kev x - ray photons per unit dark matter target of @xmath39 considering a galaxy similar to the milky way , and taking a total mass of @xmath40 , this leads to an x - ray luminosity of about @xmath41 which appears to be too small to be detectable if compared , for example , to the typical unresolved x - ray emission for a milky way - type galaxy of @xmath42 notice that the effective target number density is likely over - estimated in eq . ( [ eq : lum ] ) , as one should really only consider dark matter particles within the diffusion region where galactic cosmic rays are confined . one can also compare our estimate in eq . ( [ eq : lum ] ) with the luminosity that would be produced by sterile neutrinos with a lifetime @xmath43 s and a mass of a few kev , which lie at the verge of the sensitivity of current x - ray data ( see e.g. @xcite ) . this case is relevant here since sterile neutrinos at that mass would produce exactly the same x - ray feature , a monochromatic line at an energy of a few kev . we estimate that the corresponding milky way x - ray luminosity would be on the order of @xmath44 we thus infer that the luminosity we estimate for luminous dark matter excitation from cosmic - ray scattering , eq . ( [ eq : lum ] ) , is well below the current x - ray telescope sensitivity . in what follows , we shall focus on the case of wimps , motivated by the estimates above , and by the fact that if resonances are present , those can produce interesting and potentially revealing spectral features , such as those we found in the case of scattering of dark matter off of agn jet cosmic rays @xcite . in this section , we use the analytical results of ref . @xcite on the neutralino - electron and neutralino - proton cross section with the radiation of a photon in the final state to calculate the diffuse gamma - ray galactic emission from dark matter scattering off of cosmic - ray electrons ( sec . [ sec : ele ] ) and off of cosmic - ray protons ( sec . [ sec : prot ] ) . [ sec : ele ] we want to compute the spectrum of photons coming from the scattering of cosmic rays off of the dark matter in the milky way . we shall consider the electrons first . the photon flux is given by = r__|j d_de_e , where |j & = & _ dj ( ) , + j ( ) & = & _ 0 ^ 2 r _ ds ( r(s , ) ) f(r(s , ) ) , and @xmath45 ( see e.g. ref . @xcite and references therein ) is included to take into account the fact that the cosmic ray flux is larger in the vicinity of the galactic center . in the expressions above , @xmath46 is the angle between the emitted photon and the incoming cosmic ray . to account for cosmic rays coming from every direction we integrate over the full solid angle @xmath46 . on the other hand , @xmath47 is the angle of observation from earth . we choose to focus on the direction towards the galactic center ( gc ) , where the dark matter density and the cosmic - ray density are highest ; given the typical angular resolution , for the energies of interest , of the lat @xcite , we choose an angular region of @xmath48 around the gc , which corresponds to @xmath49 . as far as the dark matter distribution @xmath50 is concerned , we use here the spherically - symmetric einasto profile ( r ) = _ s \ { - } . the cosmic ray electron flux is taken to be @xcite = k_e ( ) ^-3 , for @xmath51 between @xmath52 and @xmath53 gev . the numerical values for the various parameters we employ in the numerical evaluations of the equations above are summarized in table [ tab : param ] . .the values in this table are from ref . [ tab : param ] [ cols="<,^ , < " , ] the last ingredient is the differential cross section @xmath54 , in order to compute which , one has to specify a particle dark matter model . for the sake of illustration , we consider here the minimal supersymmetric standard model ( mssm ) , with the lightest neutralino as the dark matter candidate . -channel.__,scaledwidth=40.0% ] we recently carried out the complete calculation of the differential cross section in ref . @xcite , using the collinear approximation , _ i.e. _ taking the photon to be collinear with the final electron . we relax here , however , that assumption , while still only considering diagrams where the selectron is exchanged in the @xmath55-channel ( see fig . [ fig : diagram ] ) , since those are the only possibly resonant channels . using the same notation as in @xcite , it is useful to remind the reader about the values of the coupling between electron ( quark ) - selectron ( squark ) - neutralino : @xmath56 and @xmath57 , with @xmath58 the electric charge in units of @xmath59 , @xmath60 the standard model @xmath61 gauge coupling and @xmath62 the weinberg angle . @xmath63 and @xmath64 refer to the handedness of the electron or quark . and masses of the lightest electron - like scalar particle @xmath65.__,scaledwidth=80.0% ] in ref . @xcite , the directions of the initial electron and the final photon were fixed by the direction of the agn jet and its angle with respect to the line of sight . that made it appropriate to use the collinear approximation . here the cosmic rays do not have a fixed incoming direction , thus we want to keep the full kinematics unconstrained , and we integrate over the full solid angle . in doing so , we encounter two distinct logarithmic enhancements : ( i ) when the photon is collinear with the final electron , as in @xcite , and ( ii ) when the photon is collinear with the initial electron . notice that this integration is numerically non - trivial , as it runs over five independent variables , which can be taken to be the photon angle @xmath46 , the direction of the final state electron , and the energy of the initial state electron . we adopt here a monte carlo global adaptive integration strategy , as implemented in the mathematica package . we show our numerical results in fig . [ fig : electrons ] , where we employ , for the sake of illustration , low lightest - neutralino masses ( 10 and 50 gev ) and small separations between the mass of the lightest selectron and the lightest neutralino ( 0.5 and 1 gev ) . the mass difference sets the location of the kinematic condition that corresponds to the end of possible resonant effects . as a result , we find a sharp cutoff at 0.5 ( black line ) and at 1 gev ( red and green lines ) . notice that the drop - off is of about one order of magnitude . also , notice that the spectrum is significantly harder at energies below the kinematic drop - off , where it is slightly harder than @xmath66 . in principle , this might make this particular process distinguishable from the diffuse galactic background . also , from the standpoint of the angular distribution , this process is _ not _ isotropic , as the corresponding intensity correlates with the product of the cosmic - ray density times the dark matter density . finally , we note that the level of intensity we predict for the signal for the models under consideration is only 3 - 4 orders of magnitude smaller than the isotropic diffuse background @xcite . while a full - scale simulation of the detectability with fermi lat of the signal we calculated here is beyond the scope of the present analysis , we argue that the level of the gamma - ray emission predicted e.g. in the models illustrated in fig . [ fig : electrons ] might be potentially detectable . at a gamma - ray energy of around 1 gev , in the same units as in fig . [ fig : electrons ] , the extragalactic diffuse emission is at the level of @xmath67 , and has a very regular power - law spectrum @xcite . considering an angular region of interest of 1 sr , the lat effective area of @xmath68 at 1 gev @xcite , and an observing time of 3 years , @xmath69 s , this corresponds to a total number of isotropic background photons collected at energies around 1 gev on the order of @xmath70 , and a total number of signal events ( for e.g. the red line model of fig.[fig : electrons ] ) of around @xmath71 . this would correspond to a relatively low signal - to - noise ( for example , @xmath720.3 ) , but in principle an excess corresponding to the sharp spectral feature we predict here could be detectable with fermi lat . a further handle would be to correlate the detected , albeit with low statistical significance , bump with its angular distribution , that should trace the product of the dark matter times galactic cosmic - ray density . [ sec : prot ] in the case of protons , the calculation is carried out at the parton level first , with a subsequent integral over the parton distribution functions ( pdfs ) @xmath73 . we thus consider = r__|j d_de_p dx _ i = u , d f_i(x ) ( e_p , x ) , where the flux of cosmic ray protons = k_p ( ) ^-2.7 . the calculation is analogous to the one done in the electron case . for numerical purposes , it is convenient to make the change of variables @xmath74 , with @xmath75 the energy of the incoming quark . doing so , there is no dependence on @xmath76 in the differential cross section and the integrals over @xmath76 and @xmath75 can be done separately . and masses of the lightest quark - like scalar particle @xmath77.__,scaledwidth=80.0% ] as we see in table [ tab : param ] , the normalization for the cosmic - ay proton flux is two orders of magnitude higher than the one for electrons . naively , one might thus hope to get a higher photon signal , but this is not actually the case . the coupling constants enter the cross section as @xmath78 , and because of the fractional electric charge of the quarks , almost two orders of magnitude are lost there . this is of course only a simple model - dependent accident . the further integration over the pdfs , however , also suppresses the signal , because having partons at a given energy requires having much larger proton energies , and the spectrum of cosmic - ray protons falls very steeply with energy . as a result , we find that the photon flux coming from proton scattering ends up being comparable to the flux from electron scattering for the particular case of supersymmetric dark matter . using again for illustrative purposes the case of supersymmetry , we choose three sets of lightest neutralino - lightest squark mass pairs in fig . [ fig : protons ] : @xmath79 gev . notice that , due to the integration over the parton distribution functions , the sharp drop - off we found for the cosmic - ray electron case is now smeared out over a broader range of photon energies , but is still noticeable . the overall normalization is still in a range which is , for some models , within 3 - 4 orders of magnitude of the isotropic gamma - ray background . notice that we find for the proton case a kinematical condition on the minimum quark energy which is completely analogous to the electron case @xcite e_q^min = , from which we see that the final photon energy can not exceed @xmath80 . that explains , for example , the sudden drop to zero at around 5 gev for the line with @xmath81 gev in fig . [ fig : protons ] . we studied the process of galactic cosmic - ray electron and proton scattering off of dark matter with the emission of a photon in the final state . we argued that the gamma - ray flux from this process is , for wimp models , suppressed compared to gamma rays from dark matter pair - annihilation . however , if the latter is in turn suppressed , or nonexistent as in asymmetric dark matter models , cosmic - ray - dark matter scattering can offer a unique opportunity for indirect dark matter searches with gamma rays . we find that , while for dipolar and luminous dark matter the intensity of the signal is well below detectability , for wimp models resonances can enhance the signal to a detectable level in principle . we illustrated the expected photon spectrum for the particular case of supersymmetric dark matter , and for a few particle dark matter and selectron / squark masses . we found that the resonant structure entails a sharp spectral drop - off , at energies corresponding to the mass difference between the selectron / squark and the lightest neutralino . the spectrum is harder than @xmath82 below the cutoff , and softer , and much more suppressed , at higher energies . in conclusion , while this process might be sub - dominant with respect to traditional searches for pair - annihilation signatures , it might still offer a unique chance of detecting indirectly dark matter with gamma - ray telescopes . the authors gratefully acknowledge conversations with tom banks , john march - russell and tomer volansky . sp and lu are supported in part by an outstanding junior investigator award from the us department of energy and by contract de - fg02 - 04er41268 ( to sp ) .
we consider the process of scattering of galactic cosmic - ray electrons and protons off of dark matter with the radiation of a final - state photon . this process provides a novel way to search for galactic dark matter with gamma rays . we argue that for a generic weakly interacting massive particle , barring effects such as co - annihilation or a velocity - dependent cross section , the gamma - ray emission from cosmic - ray scattering off of dark matter is typically smaller than that from dark matter pair - annihilation . however , if dark matter particles _ can not _ pair - annihilate , as is the case for example in asymmetric dark matter scenarios , cosmic - ray scattering with final state photon emission provides a unique window to detect a signal from dark matter with gamma rays . we estimate the expected flux level and its spectral features for a generic supersymmetric setup , and we also discuss dipolar and luminous dark matter . we show that in some cases the gamma - ray emission might be large enough to be detectable with the fermi large area telescope .
over the past few years an increasing interest to the motion of living organisms of different size and complexity ( movement ecology @xcite ) has lead to a large number of new experiments , some of them based on quite elaborated tools , allowing for study of this motion and for description of its observed trajectories . examples are the works on dicstyostelium @xcite , also under influence of chemotactic stimulus @xcite , on human motile keratinocytes and fibroblasts @xcite , on motile pieces of physarum @xcite , on flagellate eukaryotes ( euglena ) @xcite , also in time dependent light fields @xcite , and on more complex organisms as gastropods @xcite , zooplankton @xcite , birds @xcite and zebra - tail fish @xcite . one has also considered motion in the presence of boundaries which might substantially change its effective properties @xcite . similar conceptual approaches were used to describe the motion of non - living motile objects , so called self - propelled particles , exhibiting similar trajectories @xcite . theoretical modeling of self - motile objects goes back to the beginning of the last century @xcite . trajectories often appear to be stochastic @xcite , thus special attention has been paid to the development of different stochastic models for various physical or biological situations . these investigations yielded a great variety of possible mathematical models for the description of motion of self - propelled objects . examples are the discrete hopping with a given turning angle distribution @xcite , the run and tumble model @xcite with polynomial waiting time densities exhibiting superdiffusive behavior etc . in order to determine the respective mean squared displacement ( msd ) of the diffusive motion langevin equations with associated kinetic equations for the probability densities have been introduced to describe propulsive motion under influence of noise sources , as modeled by white gaussian noise @xcite , by gaussian ornstein uhlenbeck process @xcite , or by dichotomic markovian process @xcite . escape rates of active particles @xcite , of active particles in external fields @xcite and active transport in cells @xcite have been studied as well . crawling patterns of drosophila larvae motion where analyzed in @xcite by using a bimodal persistent random walk model . here , the four parameters of the corresponding model were proved to be distinct for larvae with specific genetic mutations . this work elucidates the usefulness of elaborate , multiparametric models . the origin of stochasticity , i.e. the nature of the noise sources , might be quite different , e.g. they may be due to neuronal activity , or due to interactions with the environment or with neighbors . the patchiness of the distribution of prey @xcite can be also considered as a noise source . there is still a debate on how to correctly describe various patterns of animal motion @xcite . in the present work we extend the model of an active particle moving at a constant speed with a stochastic angular dynamics @xcite to a more general case , when the random torques are described by an @xmath1-stable noise . these lvy or @xmath1-stable noise sources lead to a directed motion interrupted by large jumps in the angle of orientation @xcite . the sudden changes in the direction of movement and also the longer periods of curling appearing in this model might remind in particular the run and tumble processes @xcite . in what follows we discuss two different angular dynamics . in the first case , [ sec_tor_white ] , we consider the motion in the presence of a constant torque and of a source of a symmetric @xmath1-stable white noise . afterwards , in sec . [ sec_oup ] , we look at a non - white noise as generated by an ornstein - uhlenbeck process ( oup ) driven by a cauchy noise source . for both dynamics we derive the analytical expressions for the mean squared displacement ( msd ) and for the effective diffusion coefficient , and discuss the crossover from the ballistic motion to normal diffusion in coordinate space . we show that the msd , the diffusion coefficient , and the crossover time do not depend on the correlations in the noise source . recently , experiments on beads diffusing on macromolecules were performed @xcite and caught attention . in these experiments single particle tracking was used to follow the trajectories of the beads . the beads performed ( passive ) fickian diffusion , but a non - gaussian distribution of displacements was observed . these observations point out the necessity to pay special attention to the moments of the displacement distributions higher than the msd . in our model indeed , at difference to the behavior of the msd , the higher moments of the displacement are sensitive to correlations . in particular , in the diffusive regime of the oup - driven active particles we find an exponential distribution of displacements at short time lags while at longer time lags this distribution tends to a gaussian , an observation similar to the one reported by wang et al . @xcite . in our case this effect is caused by correlations in the noise which create an additional persistence in the motion . the crossover time after which the gaussian distribution is established depends strongly on these correlations . the the time evolution of the position vector @xmath5 of the particle is described by the following set of equations : @xmath6 @xmath7 where the angle @xmath8 gives the orientation of the velocity vector whose absolute value is @xmath9 . @xmath10 is a constant torque . the noise @xmath11 considered here is a @xmath1-stable noise . according to ditlevsen @xcite and schertzer @xcite , the associated fokker - planck - equation for such a system is given by : @xmath12 herein @xmath13 stands for the @xmath1-th symmetric riesz - weyl fractional derivative . the case @xmath14 corresponds to a gaussian white noise . the parameter @xmath1 controls the sudden large changes in the velocity orientation . for lower values of @xmath1 the noise distribution becomes sharper around the center and the tails become more pronounced , so that the probability of large sudden changes in @xmath15 increases , while small changes become less likely . the conditional probability density of the orientation @xmath15 at time @xmath16 , given the initial angle @xmath17 at time @xmath18 of the velocity vector is given by the following expression @xcite : @xmath19 in the following we will set the initial values to @xmath20 and @xmath21 . with the help of eq . the msd for this dynamics can be easily calculated using the green - kubo relation . for simplicity of notation , we move the origin of the coordinate system to the initial position of the particle at time @xmath18 . hence we have as as initial values @xmath22 or @xmath23 and @xmath24 , if not explicitly defined otherwise . we assign @xmath25 . the msd becomes : @xmath26 introducing @xmath27 and @xmath28 leads to the following form of the msd : @xmath29 for short times @xmath16 the msd shows ballistic behavior , i.e. @xmath30 . for times @xmath31 with @xmath32 being the crossover time , the motion becomes diffusive . the persistence length of the motion is @xmath33 and by that dependent on the noise type through the parameter @xmath1 . the effective diffusion coefficient @xmath34 is given by @xmath35 as a function of the parameters of the noise : simulation ( symbols ) and theory ( curves ) for @xmath36 . ] in the case of gaussian white noise @xmath37 the result of weber et al.@xcite is recovered . for @xmath38 the effective diffusion coefficient @xmath39 decays with the growth of the scale parameter @xmath40 as a power law , with a power given by the index of the lvy distribution . using eq . we see that the effective diffusion coefficients becomes @xmath41 . in fig.[msd_anf ] we show the effective diffusion coefficient @xmath34 for various values of @xmath1 and non - zero torque , as obtained in simulations and from ( eq . ) . the effective diffusion coefficient always shows a maximum at @xmath42 , where it is equal to @xmath43 . this maximal value of the diffusion coefficient does not depend on the properties of the noise . for @xmath44 the diffusion coefficient vanishes for @xmath45 and for @xmath46 . how fast it vanishes depends on @xmath1 . the higher the value of @xmath1 the smaller becomes the width of the peak . all curves intersect at @xmath47 . one also notices that the diffusion for @xmath48 for a given @xmath40 can be both faster or slower than in the gaussian case . for small noise intensities the motion of the particle is dominated by the torque , the particle moves in circles and the diffusion is slow . increasing the noise stretches the trajectories till the diffusion reaches the maximum . behind the maximum the motion becomes noise - dominated , and the diffusion coefficient falls again . simulations show that for times @xmath16 larger than the crossover time @xmath49 the transition probability density of the spatial process becomes gaussian , @xmath50 notably , as simulations show , the crossover time between ballistic and diffusive behavior coincides with the establishment of this gaussian displacement distribution . the distribution of the final position depends only on the distance between the final and the initial points , and the dependence on the angle of the vector @xmath51 with respect to the initial orientation of the velocity is lost . as we had moved the origin of the coordinate system to the position of the particle at time @xmath18 ( @xmath52 ) , we write @xmath53 instead of @xmath54 for the absolute displacement . we set @xmath20 . the statistics of the absolute displacements is given by the rayleigh distribution . the probability density of @xmath55 reads : @xmath56 where we omit @xmath20 . we now consider colored noise in the angular dynamics . we use the ornstein - uhlenbeck process ( oup ) with a cauchy noise ( @xmath57 ) for the torque . the equations for the spatial coordinates and for the time evolution of the velocity ( eq . ) remain the same as before . the angular dynamics is given by @xmath58 where @xmath11 is white noise with increments following the cauchy distribution . in the limit @xmath59 in eq . this model coincides with the model from the previous section for @xmath57 . in the limit @xmath60 the noise vanishes , but if one takes @xmath61 another interesting limiting case emerges : now the angle @xmath62 becomes a lvy process : since in this limit @xmath63 we get @xmath64 ( here we took @xmath65 ) . by introducing the oup , we get a stronger persistence in the particle s motion and introduce a new time scale , the correlation time @xmath2 . for times larger the crossover time @xmath0 the particles again perform a diffusive motion . surprisingly , as we proceed to show , the expressions for the crossover time , the msd and the diffusion coefficient coincide with eqs . and of the previous section . the onset of the asymptotic gaussian displacement distribution is however characterized by a new time scale @xmath4 which is in general different from @xmath0 as given by eq .. the fokker - planck equation the oup with cauchy noise , eq.([oup ] ) , has the following form : @xmath66 the transition probability for eq.([fok_theta ] ) , given the initial condition @xmath67 at time @xmath18 , was first obtained in ref . @xcite : @xmath68 the width of the stationary cauchy distribution for @xmath69 depends on @xmath40 only . for @xmath59 eq.([fok_phi ] ) becomes the white noise limit ( eq.([fok_white_phi ] ) ) . the angular transition probability density was determined in ref.@xcite : @xmath70 with @xmath71 and @xmath67 , @xmath17 being the values of @xmath69 and @xmath15 at time @xmath18 . the angular variable @xmath15 is a deterministic functional ( integral ) of the stochastic variable @xmath69 , and in @xcite it is pointed out that for fixed initial @xmath67 eq . represents a markov process . the cauchy case is the only case of an oup with an @xmath1-stable noise where the integrated process is markovian , and therefore the simplest one for the further analytical treatment . adding the torque and setting the initial time @xmath18 to zero results in @xmath72 using eq . one easily derives the expression for the msd . since the msd does not depend on the initial angle , we set @xmath21 . the second initial value @xmath67 has been averaged over the stationary limit of eq.([ptheta ] ) which is achieved for @xmath73 . the msd for the oup - cauchy is equal to that of an active particle from the previous sec . [ sec_tor_white ] , eqs . and with @xmath57 . averaging over initial conditions cancels any additional time dependence which would reflect the correlation behind the evolution of the direction of motion , in contrast to the oup with gaussian white noise @xcite . as the msd in our correlated case is the same as for an active particle with white angular noise with @xmath57 , the following properties stay the same : for times @xmath74 with @xmath75 the active particle moves ballistic . for times @xmath31 the regime of normal or fickian diffusion sets on , with the diffusion coefficient given by eq . with @xmath57 . the crossover from ballistic to diffusive behavior is well seen in fig.[figoupmsd ] . the behavior of the effective diffusion coefficient in dependence on other parameters was already shown in fig.[msd_anf ] for @xmath57 . for non - vanishing torque this diffusion coefficient has a maximum at @xmath76 and tends to zero for large and for small noise intensities . , @xmath77 , @xmath38 , black line corresponds to eq . with @xmath57 , theory and simulation align perfectly , the msd does not depend on the correlation time @xmath2 the crossover time from ballistic to diffusive motion is indicated by dashed line at time @xmath0 . ] it is remarkable that the msd and the diffusion coefficient are independent of the correlation time @xmath2 , so that the underlying correlations can not be detected from measuring the msd . therefore one might expect to observe a gaussian distribution for the displacements , eq . , for times @xmath78 and , respectively , the rayleigh distribution , eq . , for the absolute displacement for all such times . this expectation is however wrong . as we show in the following plots , the probability density of the absolute displacement @xmath79 deviates strongly from the rayleigh distribution even for times well beyond the crossover time to diffusive behavior @xmath0 . depending on @xmath2 , the asymptotic convergence to the expected gaussian and rayleigh distributions which are independent of @xmath2 may take place only at times which are considerably larger . at @xmath80 for given correlation time @xmath2 . black line corresponds to the absolute displacement distribution eq . at the given time @xmath16 , with @xmath81 from eq.([diff]),(@xmath57 ) . the black dashed line shows the exponential distribution from eq . with @xmath81 from eq . , with @xmath57 at the given time . ; b ) : the logarithmic plot of the time evolution displacement distributions rescaled by @xmath82 for indicated times and parameters . colored symbols correspond to simulations . the black line corresponds to eq . for ( @xmath83 ) . the black dashed line shows the exponential distribution from eq . with @xmath81 from eq.,(@xmath57 ) at the given time . at time @xmath84 the displacement distribution ( blue circles ) displays an exponential behavior and at time @xmath85 ( red triangles ) it has come closer to a gaussian.,title="fig : " ] at @xmath80 for given correlation time @xmath2 . black line corresponds to the absolute displacement distribution eq . at the given time @xmath16 , with @xmath81 from eq.([diff]),(@xmath57 ) . the black dashed line shows the exponential distribution from eq . with @xmath81 from eq . , with @xmath57 at the given time . ; b ) : the logarithmic plot of the time evolution displacement distributions rescaled by @xmath82 for indicated times and parameters . colored symbols correspond to simulations . the black line corresponds to eq . for ( @xmath83 ) . the black dashed line shows the exponential distribution from eq . with @xmath81 from eq.,(@xmath57 ) at the given time . at time @xmath84 the displacement distribution ( blue circles ) displays an exponential behavior and at time @xmath85 ( red triangles ) it has come closer to a gaussian.,title="fig : " ] in fig . [ figspat]a the absolute displacement distributions rescaled by @xmath82 for three different correlation times @xmath2 at time @xmath86 are shown . the chosen time is well beyond the crossover time @xmath0 to the diffusion regime , i.e. @xmath31 , and also much larger than correlation time , @xmath87 . all three distributions are different . only for the shortest correlation time the simulation results and the prediction of eq . ( the black line in fig . [ figspat]a ) coincide . for longer correlation times the absolute displacement shows an exponential behavior ( black dashed line , eq . ) . such behavior is not observed in simulations of particles driven by an oup with gaussian white noise . in fig . [ figspat]b the time evolution of the absolute displacement is presented . the results for three different time lags , all within the diffusion regime , are plotted . the black line corresponds to the result of eq . at time lag t=49.9 , with the effective diffusion coefficient according to eq . with @xmath57 . as the simulations show , the distribution ( red triangles in fig.([figspat]b ) ) at this time has still not fully approached its asymptotic shape ( black line ) . at smaller time lags the displacements clearly deviate from this line , exhibiting an exponential behavior . the black dashed line shows such an exponential decay . for various intensities of the cauchy distributed noise in eq .. colored symbols correspond to simulations . black line is the distribution from eq .. black dashed lines are exponentials from eq .. b ) : simulation results for the kurtosis of the distributions for different correlation times . the values approach asymptotically the gaussian limit ( dashed black line ) indicating the crossover to a gaussian displacement distribution.,title="fig : " ] for various intensities of the cauchy distributed noise in eq .. colored symbols correspond to simulations . black line is the distribution from eq .. black dashed lines are exponentials from eq .. b ) : simulation results for the kurtosis of the distributions for different correlation times . the values approach asymptotically the gaussian limit ( dashed black line ) indicating the crossover to a gaussian displacement distribution.,title="fig : " ] fig . [ figth]a shows the the absolute displacement distribution for different values of the noise strength @xmath40 . colored symbols stand for simulations . the black line shows the rayleigh distribution from eq . rescaled by the distance @xmath55 . the black dashed lines are exponentials from eq . ) . for higher noise intensities the distributions decay faster . higher noise intensities correspond to an increase in directional changes which then lead to smaller spatial increments in a given time interval . it is in principle possible ( but tedious ) to calculate higher moments of the displacement using eq .. we choose here a phenomenological approach . we make the following exponential ansatz for the normalized distribution density of the displacements with the time dependent characteristic length scale @xmath88 : @xmath89 in agreement with our initial condition , this scale at initial time @xmath90 should vanish , i.e. @xmath91 . for the absolute displacement the probability distribution reads : @xmath92 we require that the expectation value @xmath93 for the squared absolute displacement ( eq.([eqn_displ_exp ] ) ) is equal to the long time limit of the calculated msd ( eq.([msd ] ) ) @xmath94 on the other hand , it results from eq.([eqn_displ_exp ] ) that @xmath95 . therefore , the absolute displacement distribution @xmath79 reads : @xmath96 the black dashed lines in fig.([figspat ] ) and fig.([figth]a ) correspond to eq.([eqn_displ_exp_f ] ) . this distribution fits the simulations very well for times where the displacement is not yet gaussian , although eq.([eqn_displ_exp_f ] ) does not depend on the correlation time . [ figth]b shows simulations for the fourth moment of the displacement divided by the squared msd ( kurtosis ) . in two dimensions a value of two corresponds to a gaussian distribution , shown by the horizontal black dashed line in fig.([figth]b ) . a higher value of the kurtosis indicates heavier tails and sharper central peak in the distribution , while a smaller value corresponds to lighter tails and flatter peak . in experiments it can be calculated from the particle positions at different instances of time . in fig . [ figth]b the kurtosis for three different correlation times @xmath2 is plotted . with growing time asymptotically the gaussian limit is reached for all correlation times . for short correlation times @xmath97 the limit is reached from below , while for larger @xmath2 the kurtosis exhibits a maximum and then approaches the limit from above , indicating the observed transient exponential or heavy - tailed regime . the height of the maximum grows with increasing correlation time , so the deviation from the gaussian displacement becomes more pronounced in the transient regime . with increasing @xmath2 the time till the asymptotics is reached also increases . we give here an estimate of the crossover time @xmath4 from non - gaussian to gaussian displacement distributions . we require , for @xmath98 , that the msd equals the squared correlation length @xmath99 , a typical length of the trajectory within the correlation time @xmath100 of the noise , i.e. @xmath101 note that the length @xmath102 differs from the persistence length , as introduced in eq . , being the characteristic scale of transition from ballistic to diffusive motion . since the particle is already in the diffusive regime the msd becomes @xmath103 and in consequence we get : @xmath104 for zero torque @xmath38 in eq.([diff ] ) @xmath57 the estimate for the crossover time becomes : @xmath105 comparing this result for @xmath4 with fig . [ figspat]b indicates that this time @xmath4 is a lower bound for the establishment of the rayleigh displacement distribution from eq .. hence , for @xmath106 the distribution of displacements becomes gaussian . ( black ) and @xmath107 ( blue ) , right : @xmath108 ( black ) and @xmath109 ( red ) , colored trajectories shifted for better visibility , title="fig : " ] ( black ) and @xmath107 ( blue ) , right : @xmath108 ( black ) and @xmath109 ( red ) , colored trajectories shifted for better visibility , title="fig : " ] at the end of this section we show a typical trajectory of the process and discuss the importance of sampling time lags . fig.([figoup ] ) shows always the same sample trajectory for @xmath110 and @xmath111 with different sampling times ( black : the sample time lag @xmath108 , blue : @xmath107 , red : @xmath109 ) . for the black trajectory positions every @xmath108 where plotted and connected by a straight line . for the other sampling time lags the same procedure was done accordingly . for the small sampling time lag ( black line ) one sees a smooth structure with spirals or curls reminding of a run and tumble motion . the curly structure is influenced by the correlation time @xmath2 . higher correlation time @xmath2 increases length and size of a spiral . setting the correlation time close to zero @xmath112 removes the curls and makes the trajectory less smooth ( not shown ) . the blue trajectory in fig.([figoup ] ) is the same trajectory as the black one but sampled differently ( @xmath107 ) ; the sampling time lag @xmath107 still belongs to the lags at which non - gaussian displacement distributions are observed . it still contains some information about the curls . ignoring the underlying curly structure one might interpret the blue trajectory as one where a particle spends some time in a caged area or is rather immobile ( maybe reorients itself ) and then moves in rather straight stretches . the red trajectory corresponds to sampling where the gaussian displacement distribution of displacements during the time lags is established . there are no remnants of the spiral structure visible . we studied the stochastic dynamics of active particles moving at a constant speed whose direction of motion is influenced by an @xmath1-stable noise source and by a constant torque . first , the noise was considered to be white , later on , we looked at a colored cauchy noise as generated by an ornstein - uhlenbeck process with the characteristic time @xmath2 . the model with white noise generates the motion showing a crossover from the ballistic to the diffusion behavior similar to a behavior for the gaussian case . the behavior under the colored noise generates the pattern of motion strongly resembling the run and tumble situations or the behavior seen in experiments like @xcite where particles spend some time in confined motion , reorient themselves and then move on . for both physical models we derived analytical expressions for the mean squared displacement and for the effective diffusion coefficient astonishingly , in the case of cauchy - noise , the existence of correlations in the noise does not influence the msd , the effective diffusion coefficient and the crossover time @xmath0 from ballistic to diffusive motion coincide for both models with @xmath57 . the distribution function of the displacements is however strongly influenced by correlations in the noise even at the times well inside the diffusive regime . in particular , results of numeric simulations for times larger @xmath0 can be well fitted by an exponential distribution . this exponential behavior appears to be a transient . a crossover to gaussian distribution of displacements takes place at a third characteristic time @xmath113 , connected with @xmath2 and @xmath0 . this colored noise mechanism might offer an approach to describe observations of a transient non - gaussian displacement distribution in diffusion experiments like those performed by wang et al . @xcite . the observed effect is absent in the case of the oup with gaussian white noise source , i.e with @xmath14 , and becomes the more pronounced , the smaller is the value of @xmath1 . generally , our findings underline the importance to investigate the behavior of the higher moments of displacement both in experiments like @xcite and in simulations . these moments might provide new information on the persistence of the motion . the studied case of the oup with cauchy noise is special , and the independence of msd of the correlation time @xmath2 of the noise is not a general situation . nevertheless , we expect deviations from a gaussian distribution generally in models with a correlated non - gaussian noise in the angular dynamics , like in @xcite . this work was developed within the scope of the irtg 1740 funded by dfg / fapesp . lutz schimansky - geier acknowledges support from the humboldt - university at berlin within the framework of excellence initiative ( dfg ) . the authors thank dr . b. dybiec ( krakow ) for fruitful discussions . xxxxxxxxxxxxxxxxxxx r. natan , proc . 105 , 19050 ( 2008 ) . bdecker , c. beta , t.d . frank , and e. bodenschatz , epl 90 , 28005 ( 2010 ) . g. amselem , m. theves , a. bae , e. bodenschatz , and c. beta , plos one 7 . e37213 ( 2012 ) . d. selmeczi , s. mosler , p. h. hagedorn , n. b. larsen , and h. flyvbjerg , biophys . j. 89 , 912 ( 2005 ) . b. rodiek and m.j.b . hauser , europ . journ.- special topics , 223 , 1199 ( 2015 ) . p. romanczuk , m. romensky , d. scholz , v. lobaskin , and l. schimansky - geier , europ . journ.- special topics , 223 , 1215 ( 2015 ) . m. romensky , d. scholz , and v. lobaskin , j. royal soc . interface 12 , 20150015 ( 2015 ) . l. seuront , a .- c . duponchel , and c. chapperon , j. physica a 385 , 573 , ( 2007 ) . a. ordemann , g. balazsi , and f. moss , physica a 325,260 ( 2003 ) . a.m. edwards , r.a . phillips , n.w . watkins , m.p . freeman , e.j . murphy , v. afanasyev , s.v . buldyrev , m.g.e . da luz , e.p . raposo , h.e . stanley , and g.m . viswanathan , nature 449 , 1044 ( 2007 ) . j. gautrais , c. jost , m. soria , a. campo , s. motsch , r. fournier , s. blanco , and g. theraulaz , j. math . biol . 58 , 429 ( 2009 ) . s. van teeffelen and h. lwen , phys . e * 78 * , 020101 ( 2008 ) . s. van teeffelen , u. zimmermann , and h. lwen , soft matter * 5 * , 4510 ( 2009 ) . h. ke , s. ye , r. lloyd - calroll , and k. showalter , j. phys . chem . a 114 , 5462 ( 2010 ) . d takagi , a.b . braunschweig , j. zhang , and m.j . shelley , phys . 110 , 038301 ( 2013 ) . b ten hagen , f. kummel , d. takagi , h. lwen , and c. bechinger , nature communications 5 , 4829 ( 2014 ) . k. pearson , _ a mathematical theory of random migration _ , dulau and co. , london , 1906 r. frth , zeitschrift fr physik 2 , 244 ( 1920 ) . h.c.berg , phys . today , 53 , 24 ( 2000 ) . d. campos , v. mendez , and i. llopis , j. theor . 267 , 526 ( 2010 ) . p. romanczuk , m. br , w. ebeling , b. lindner , and l. schimansky - geier eur . phys . j. spec . 202 , 1 ( 2012 ) . l. schimansky - geier , u. erdmann , and n. komin , physica a * 351 * , 51 ( 2005 ) . l. haeggqwist , l. schimansky - geier , i.m . sokolov , and f. moss , eur . j. st * 157 * , 33 ( 2008 ) . f. thiel , l. schimansky - geier , and i.m . sokolov , phys . rev.e 86 , 021117 ( 2012 ) . a. s. mikhailov and d. meinkhn , stochastic dynamics , edited by l. schimansky - geier and t. pschel ( springer , berlin , 1997 ) , p. 334 . w. blum , w. riegler , and l. rolandi , particle detection with drift chambers , 2nd ed . ( springer , berlin,2008 ) . f. peruani and l.g . morelli , phys . 99 , 10602 ( 2007 ) . p. degond and s. motsch , j. stat . phys . 131 , 989 ( 2008 ) . c. weber , p.k . radtke , l. schimansky - geier , and p. hnggi phys.rev.e 84 , 011132 ( 2011 ) . c. weber , i.m . sokolov , and l. schimansky - geier , phys . e 85 , 052101 ( 2012 ) . burada and b. lindner phys . rev.e 85 , 032102 ( 2012 ) . w. ebeling , l. schimansky - geier , a. neiman , and a. scharnhorst , fluct . noise lett . 5 , l185 ( 2005 ) . a. geiseler , p. h nggi , and g. schmid , eur . j. b 89 , 17 ( 2016 ) . a. godec and r. metzler j. phys . a 49 , 364001 ( 2016 ) . a. reynolds , ecology 89(8 ) , 2347 ( 2008 ) . gnther , g. nettesheim , and g.t . shubeita , scientific reports 6 , 27972 . b. wang , s. m. anthony , s.c . bae , and s. garnick , proc . usa 106 , 15160 - 15164 ( 2009 ) . b. wang , j. kuo , s.c . bae , and s. granick , nature materials 11 , 481 ( 2012 ) . b. dybiec , e. gudowska - nowak , and p. hnggi , phys . e 73 , 046104 ( 2006 ) . ditlevsen , phys . rev.e 60 , 172 ( 1999 ) . d. schertzer , m. larchevque , j. duan , v.v . yanowsky , and s. lovejoy , j. math . 42 , 200 ( 2001 ) . b. j. west and v. seshadri , physica a 113 , 203 ( 1982 ) . p. garbaczewski and r. olkiewicz , j. math . 41 , 6843 ( 2000 ) . uhlenbeck and l.s . ornstein , phys . 36 , 823 ( 1930 ) .
we investigate the stochastic dynamics of an active particle moving at a constant speed under the influence of a fluctuating torque . in our model the angular velocity is generated by a constant torque and random fluctuations described as a lvy - stable noise . two situations are investigated . first , we study white lvy noise where the constant speed and the angular noise generate a persistent motion characterized by the persistence time @xmath0 . at this time scale the crossover from ballistic to normal diffusive behavior is observed . the corresponding diffusion coefficient can be obtained analytically for the whole class of symmetric @xmath1-stable noises . as typical for models with noise - driven angular dynamics , the diffusion coefficient depends non - monotonously on the angular noise intensity . as second example , we study angular noise as described by an ornstein - uhlenbeck process with correlation time @xmath2 driven by the cauchy white noise . we discuss the asymptotic diffusive properties of this model and obtain the same analytical expression for the diffusion coefficient as in the first case which is thus independent on @xmath2 . remarkably , for @xmath3 the crossover from a non - gaussian to a gaussian distribution of displacements takes place at a time @xmath4 which can be considerably larger than the persistence time @xmath0 .
electron scattering is the preferred tool for extracting information on the spatial and momentum distribution of the quarks in nucleons . high energy scattering provides a clean picture of the quarks momentum distribution for a nucleon boosted into infinite momentum frame ( imf ) . measurements of elastic scattering at lower energy scales allow extraction of the nucleon form factors , which can be related to the spatial distribution of charge in the rest frame of the nucleon . however , obtaining rest frame charge distributions requires model - dependent relativistic boost corrections , limiting our ability to extract these distributions from data . significant work has gone into better understanding the issues involved in studying nucleon distributions , as well as providing unified descriptions of the space and momentum distributions of the quarks . the present discussion and the papers on which it is based would not have been possible without the great amount of experimental technique , effort and ingenuity that has been used recently to measure the electromagnetic form factors of the nucleon@xcite . these quantities are probability amplitudes that the nucleon can absorb a given amount of momentum and remain in the ground state , and are related to the nucleon charge and magnetization densities . we note that there will be an institute for nuclear theory program held in the fall of 2009 that is devoted to the electromagnetic physics of the jefferson laboratory upgrade to 12 gev . please see the website : www.int.washington.edu/programs/09-03.html . we begin our analysis by reviewing recent work@xcite that determines the transverse charge density of the neutron in a model independent way . this work surprisingly found that the central charge density is negative . then we discuss how the inclusive - exclusive connection is used to provide an interpretation for this fact@xcite . we begin by by presenting definitions of the form factors . let @xmath3 be the electromagnetic current operator , in units of the proton charge . then the nucleon form factors are given by the matrix element p,| j^(0)| p,=|u(p, ) ( ^f_1(q^2)+iq_f_2(q^2 ) ) u(p,),where @xmath4 is the nucleon mass , and the momentum transfer @xmath5 is taken as space - like , so that @xmath6 the nucleon polarization states are chosen to be those of definite light - cone helicities @xmath7.@xcite the charge ( dirac ) form factor is @xmath8 , normalized such that @xmath9 is the nucleon charge , and the magnetic ( pauli ) form factor is @xmath10 , normalized such that @xmath11 is the anomalous magnetic moment . the sachs form factors are @xmath12 in the breit frame , in which @xmath13 , @xmath14 is the nucleon helicity flip matrix element of @xmath15 . this has been interpreted as meaning that @xmath14 is the three - dimensional fourier transform of the charge density in the rest frame . indeed , the scattering of neutrons from the electron cloud of atoms measures the derivative @xmath16 at @xmath17 , which has been widely interpreted as the mean - square charge radius of the neutron . however , a direct probability or density interpretation of @xmath14 is spoiled by a non - zero value of @xmath18 , no matter how small . this is because relativistic dynamics , which cause the wave functions of the initial and final nucleons of different momenta to differ , must be used . the final wave function is related to the initial one by a complicated boost operator that when acting on the initial state of a given momentum changes it to the same of the different final momentum . in general the boost operator ( or boost ) contains the effects of interactions . thus the initial and final states differ , invalidating a probability or density interpretation . it is only in the case that non - relativistic dynamics are applicable that form factors are simply the fourier transforms of the rest frame spatial distributions . it has commonly been assumed that at low momentum transfers , these corrections could be safely neglected . if we treat relativistic corrections as in atomic physics , they are governed by @xmath19 , where @xmath20 is the mass of the boosted constituent . computing the form factors involves integrating over the momentum @xmath21 so that it is replaced by the momentum transfer @xmath22 . thus for small values of @xmath18 we find g_e(q^2 ) = g_e(0)- q^26 ( d^3r r^2 ( r ) + c / m^2),where @xmath23 is an unknown coefficient . in a simple constituent quark model , @xmath24 mev , and so these @xmath25 boost corrections are negligible only for @xmath26 gev@xmath27 . however , for the neutron @xmath28 , so the corrections are expected to be relatively large . the boost correction term need not be small compared to the finite size contribution unless @xmath29 ; since both corrections scale with @xmath18 , the relative correction to the extracted radius does not vanish as @xmath18 approaches 0 . the constituent quarks represent the low @xmath18 , dressed versions of the near - massless current quarks of qcd . for current quarks of mass 510 mev , the boost corrections are important for all @xmath18 values where measurements exist . while it is possible to construct systematically improveable models which should yield a complete description of the nucleon , it is not clear how one would quantitatively determine how well any particular model fully reproduces the corrections associated with the relativistic boost . thus , there is always some model dependence in the extraction of the rest frame charge distributions from the form factors , and it is not clear how well one can quantify these corrections and uncertainties , even at very low @xmath18 . the use of light cone coordinates and the kinematic subgroup of the poincar group , which is closely related to the use of the infinite momentum frame , enables one to avoid the difficulties associated with including the boost . the basic idea is that the `` time '' variable is given by @xmath30 and the conjugate evolution operator is given by @xmath31 starting in an ordinary reference frame and making a lorentz transformation into a frame moving with nearly the speed of light in the @xmath32 direction converts the usual @xmath33 into @xmath34 . the 3 spatial variables must be different than @xmath34 , so we take @xmath35 if @xmath36 then @xmath37 and @xmath38 can be thought of as something like a @xmath39 or @xmath40 variable , but rotational invariance can not generally be used to relate the @xmath38 and @xmath41 dependence of the density . the canonically conjugate momentum to @xmath38 is @xmath42 for the transverse degrees of freedom we use the usual position @xmath43 and momentum @xmath44 variables denoted by @xmath45 where the boldface notation denotes a transverse vector . we also exploit the kinematic subgroup . that is transverse lorentz transformations , with a transverse velocity @xmath46 ( or boosts in the transverse @xmath41 or @xmath47 direction ) do not involve interactions . in particular , these transformations are defined by @xmath48 with @xmath49 changed so that @xmath50 is not changed . these transformations are just like the non - relativistic galilei transformation except that @xmath51 appears instead of a mass . this means that we are allowed to do fourier transformations of variables provided transverse degrees of freedom are involved . if the momentum transfer vector @xmath52 is space - like we may use the so - called drell - yan ( dy ) frame in which @xmath53 . this means that the plus component of the momentum of the nucleon is the same before and after the absorption of the single photon . then the momentum transfer vector is in the transverse direction , so that usual two - dimensional fourier transform techniques can be used . a proper determination of a charge density requires the measurement of a density operator . we shall show that measurements of the pion form factor directly involve the three - dimensional parton charge density operator , in the infinite momentum frame imf , @xmath54 . in this frame the electromagnetic charge density @xmath15 becomes @xmath55 and _ ( x^- , ) = _ q e_q ( x^-,)^+q(x^-,)=_q e_q q^_+(x^-,)q_+(x^-,),[imfop]where @xmath56 , the independent part of the quark - field operator @xmath57 . we set the time variable , @xmath58 to zero , and do not display it in any function . the spatial structure of a hadron can be examined if one uses@xcite states that are transversely localized . the state with transverse center of mass @xmath59 set to * 0 * is formed by taking a linear superposition of states of transverse momentum : @xmath60 are light - cone helicity eigenstates @xcite and @xmath61 is a normalization factor satisfying @xmath62 . the expansion eq . ( [ eq : loc ] ) makes sense only in the infinite momentum frame @xmath63 because we must have @xmath64 . the nucleon states are normalized as @xmath65 . next we relate the charge density _ ( x^-,)= , to @xmath66 . in the dy frame no momentum is transferred in the plus - direction , so that information regarding the @xmath38 dependence of the distribution is not accessible . therefore we integrate over @xmath38 , using the relationship translational invariance , and then use our momentum expansion and the definition of @xmath8 as the helicity non - flip matrix element of @xmath55 in a drell - yan frame to find & & ( b)dx^-_(x^-,b)[rhobtx ] + & & ( b)= f_1(q^2=^2 ) e^-i= f_1(q^2)j_0(qb ) e^-i , [ rhob0]where @xmath67 is a cylindrical bessel function and @xmath68 is termed the transverse charge density , giving the charge density at a transverse position @xmath69 , integrating over all longitudinal momentum . in the infinite momentum frame , the longitudinal dimension of the nucleon is contracted to a point , and only the transverse position remains . the value of @xmath70 corresponds to the center of longitudinal - momentum of the nucleon in the transverse dimension . it is not as straightforward to isolate the magnetization distribution , and there have been multiple such densities proposed @xcite . starting with the matrix element of @xmath71 and working in the infinite momentum frame leads to the result that the transverse magnetization density to is the two - dimensional fourier transform of @xmath10 , just as the charge density is the transform of @xmath8 @xcite . this interpretation yields a difference between the magnetic and electric radii in the proton . 1.cm ( 14,10.8)(1.8,1.1 ) [ proton ] here we exploit eq . ( [ rhob0 ] ) by using recent parameterizations @xcite of measured form factors to determine @xmath68 . applying eq . ( [ rhob0 ] ) to the proton using two sets of form factor parameterizations@xcite yields the results shown in the upper panel of fig . [ fig : proton ] . the curves obtained using the two different parameterizations overlap . furthermore , there is negligible sensitivity to form factors at very high values of @xmath18 that are currently unmeasured . the density is peaked at low values of @xmath69 , but has a long positive tail , suggestive of a long - ranged , positively charged pion cloud . the neutron results for @xmath68 are shown in the lower panel of fig . [ fig : proton ] . the curves obtained using the two different parameterizations seem to overlap , but we will return to this below . the surprising result is that the central neutron charge density is negative . the negative nature of the neutron s central charge density appears to contradict two current ideas . if the neutron is sometimes a proton surrounded by a negatively charged pionic cloud , one would expect to obtain a positive central density@xcite . another mechanism involving correlations in the nucleonic wave function induced by one gluon exchange would also lead to a positive central density because the interaction between two identical @xmath1 quarks@xcite is repulsive . and transverse charge density . upper panel : @xmath8 . lower panel : @xmath72 in transverse position space . the solid curves are obtained using the fits of kelly , and the dashed curves the fits of bradford , _ et al._. reprinted with permission from ref . copyright 2007 by the american physical society.,title="fig:",width=377,height=207 ] and transverse charge density . upper panel : @xmath8 . lower panel : @xmath72 in transverse position space . the solid curves are obtained using the fits of kelly , and the dashed curves the fits of bradford , _ et al._. reprinted with permission from ref . copyright 2007 by the american physical society.,title="fig:",width=377,height=207 ] the resultant negative central density thus deserves further examination . the upper panel of fig . [ fig : f1 ] shows @xmath8 for the neutron obtained using the two different parameterizations@xcite which are observably different . however , in both cases . @xmath8 is negative for all values of @xmath18 . if @xmath8 is always negative , then taking @xmath74 in eq . ( [ rhob0 ] ) , will always yield a negative central density . the long range structure of the charge density is captured by displaying the quantity @xmath75 in the lower panel of fig . [ fig : f1 ] . at very large distances from the center , the charge density is negative , as expected in the pion cloud picture . these findings appear to contradict previous understanding of the nucleon charge distributions based on the model - dependent extraction of the rest frame charge distributions . because of the large and model - dependent boost corrections at large @xmath18 , one can not obtain information about the rest frame charge density at the very center of the nucleon from measurements of the form factors . while there is no direct experimental information on this possibility of a very small negative core in the rest frame distribution , this negative core seems to contradict the accepted explanations of the origin of the long range negative cloud which occurs along with a positively charged interior . in addition , the negative core in the imf transverse density is a feature even in models that build up the charge distribution based on a pion cloud model . it is clearly important to understand the differences between the imf charge density and the rest frame charge density to fully understand the new features of these model - independent spatial distributions . the surprising model independent result is that the density of the neutron is negative . the remainder of this presentation is concerned with trying to explain this remarkable feature of nature . generalized parton distributions ( gpds ) contain information about the longitudinal momentum fraction @xmath0 as well as the transverse position @xmath69 . information regarding the @xmath0 and @xmath69 dependence is obtained from experiment by using gpds to reproduce both deep inelastic scattering and elastic scattering data . thus we use this inclusive - exclusive connection to better understand the central neutron charge density . the widely studied gpds@xcite are of high current interest because they can be related to the total angular momentum carried by quarks in the nucleon . we consider the specific case in which the longitudinal momentum transfer @xmath76 is zero , and the initial and final nucleon helicities are identical ( @xmath77 ) . then , in the light - cone gauge , @xmath78 , the matrix element defining the gpd @xmath79 for a quark of flavor @xmath80 and zero skewness is h_q(x , t ) = p^+,,|_q(x,*0 * ) |p^+ , , , [ eq : pd ] where _ q(x,*b * ) _ + ^(-,*b * ) q_+(,*b * ) e^ixp^+x^- . [ eq : bperp ] we abbreviate @xmath81=@xmath82 and @xmath83 it is well known that gpds provide a unified description of a number of hadronic properties.@xcite of particular interest here is that for @xmath33=0 they reduce to conventional pdfs , @xmath84 , and that the integration of the charge - weighted @xmath79 over @xmath0 yields the nucleon electromagnetic form factor : f_1(t)=_q e_q dx h_q(x , t ) . [ eq : form ] the impact parameter - dependent pdf@xcite for a quark of flavor @xmath80 is the matrix element of the operator @xmath85 in the state @xmath86 : _ ^q(*b*,x ) p^+,*r*= * 0 * , | _ q(x,*b * ) |p^+,*r*= * 0 * , . [ eq : def1 ] we use the notation @xmath87 instead of the originally defined@xcite @xmath88 because this quantity is a density that gives the probability that the quark has a longitudinal momentum fraction @xmath0 and is at a transverse position @xmath47 . the quantity @xmath87 is the two - dimensional fourier transform of the gpd @xmath79 : _ ^q(*b*,x)=e^-ih_q(x , t=-^2).[ft1 ] we extract the form factor @xmath8 by integrating @xmath87 over all values of @xmath0 , multiplying by the quark charge @xmath89 , and summing over quark flavors @xmath80.@xcite the result is the imf charge density in transverse space : & ( b ) & _ q e_qdx_^q(*b*,x ) . [ rhobt ] the relations eq . ( [ rhobt ] ) and eq . ( [ rhobtx ] ) provide two expressions for the transverse density @xmath68 . the impact parameter gpd and the three - dimensional density are related by parseval s theorem . the quantity @xmath68 gives the charge density at a transverse position @xmath47 irrespective of the longitudinal momentum fraction or longitudinal position . there is a tight connection between the values of @xmath0 and the values of @xmath47 . for a given fock space component , the center of transverse momentum ( cm ) @xmath59 is given by = 0=_ix_i_i , [ tcm]where @xmath90 are the longitudinal momentum fractions and transverse positions of the @xmath91th partons . using this definition , the longitudinal momentum of a quark determines it s impact on defining the transverse position of the nucleon . if the struck quark , @xmath92 , carries nearly all of the plus component of the total momentum , then the other quarks must carry a total plus momentum of @xmath93 which approaches zero as @xmath94 . thus , @xmath95 for @xmath96 , and via eq . ( [ tcm ] ) , @xmath97 must also approach zero as long as the values of @xmath98 remain finite . thus , large values of @xmath0 correspond to a small value of @xmath69 , as the struck quark plays a large role in defining the transverse center of mass as @xmath99 . let us investigate @xmath100 to understand the origin of the neutron s negative central charge density . the quantities are not measured directly , but have been obtained from models that incorporate fits to parton distributions and electromagnetic nucleon form factors.@xcite this method exploits form factor sum rules at zero skewness to model information regarding the valence quark gpds , @xmath101 . this yields the net contribution to the form factors from quarks and anti - quarks , although it does not correspond to the valence distribution within a model for which sea distributions for quarks and antiquarks have different @xmath0 or @xmath33 dependences . the effects of strangeness are neglected in these fits . each parameterization we use@xcite incorporates the drell - yan - west@xcite relationship between the behavior of the structure function @xmath102 function near @xmath103 , measured in inclusive reactions and the behavior of the electromagnetic form factor at large values of @xmath18 , measured in the exclusive elastic scattering process . in particular , for a system of @xmath104 valence quarks , described by a power - law wave function _ x1w_2(x)=(1-x)^2n-1 _ q^2f_1(q^2)= , with the relation being that the same value of @xmath105 that defines the high-@xmath0 behavior of the structure function also defines the high-@xmath18 behavior of the form factor , thus associating the behavior of large values of @xmath0 with large momentum transfers , @xmath106 , which in turn correspond to small values of @xmath69 . so again we see we see a connection between large values of @xmath0 and small values of @xmath69 . to proceed further we use specific forms of the gpds , and these determine the details of the results . diehl _ et al._@xcite use h^q_v(x , t)=q_v(x ) , where f_q(x)=[+b_q](1-x)^3+a_qx(1-x)^2 , is the form that gives the best fit to the data . the parameter @xmath107 represents the slope of the regge trajectory , and the cteq6 pdfs@xcite are taken as input . here we use the best fit parameters , taken from the second line of table 8 of ref . these are detailed in ref . the labels @xmath80 refer to @xmath108 and @xmath1 , the @xmath108 and @xmath1 quarks in the proton . these correspond to @xmath1 and @xmath108 quarks in the neutron , if charge symmetry @xcite is upheld . for the proton , @xmath109 falls rapidly for large values of @xmath0 , which means that @xmath108 quarks dominate the parton distribution for large values of @xmath0 . for the neutron , the assumption of charge symmetry implies that @xmath110 for the neutron is the same as @xmath111 for the proton , and vice - versa . thus , the @xmath1 quarks dominate the parton distribution in the neutron for large values of @xmath0 . the distributions of ref . @xcite have @xmath112 and @xmath113(1-x).$ ] those of ref . @xcite have a more complicated form and include the additional constraint that the nucleon consists of three quarks at an initial scale of @xmath114 gev@xmath27 . we study the connection between regions of @xmath0 and regions of @xmath69 . to do this define _ ^q(b , x)_x dx e_q _ ^q(b , x ) , where @xmath89 is the quark charge in units of the proton charge ( @xmath115 ) with @xmath116 being obtained from appropriate sums of @xmath117 . this represents the contribution to the charge density from quarks in the @xmath0 region defined by @xmath118 , rather than the total density obtained by integrating over all @xmath0 . , for quarks in different @xmath118 regions : @[email protected] ( solid ) , 0.15@xmath119@[email protected] ( long - dash ) , 0.3@xmath119@[email protected] ( short - dash ) , and @[email protected] ( dotted ) . the curves have been normalized to unity at @xmath70 to emphasize the variation in width . reprinted with permission from ref . copyright 2008 by the american physical society.,width=321 ] we have pointed out above that at large @xmath0 , the struck quark plays a significant role in defining the transverse cm , so the distribution of high-@xmath0 quarks becomes localized at small values of @xmath69 . this is clearly visible in fig . [ fig : proton_scaled ] , which shows @xmath122 for different bins in @xmath118 . the curves have been scaled to yield unity at @xmath70 , to emphasize the variation in width . the four @xmath118 regions yield 58% , 25% , 14% , and 3% of the total charge , with the largest contributions coming from the bins with the smallest values of @xmath0 . for @xmath123 , the half - maximum width is 0.5 fm , while for @xmath124 , it is 0.12 fm . thus the large @xmath0 quarks ( mainly @xmath108 quarks in the proton ) play an increasingly prominent role in the charge distribution at small values of @xmath69 . the curves shown in fig . [ fig : proton_scaled ] are obtained using the gpd of ref . @xcite ; the results obtained from the guidal _ et al _ , parameterization for the gpds are barely distinguishable . the gpds of ref.@xcite also have a strong tendency to be constrained to smaller and smaller values of @xmath69 as the value of @xmath0 increases . we evaluate the gpds of all three models using the starting scale @xmath125 of each model . taking what we have learned from the proton , we now consider the neutron . figure [ fig : ud_neutron ] shows the ratio of the up- to down - quark distributions in the neutron , as extracted from various analyses of deuteron and proton data @xcite ( using different models for the nuclear corrections in dueterium ) , and from the cteq6l @xcite parameterization . for @xmath126 , the up and down quark distributions are similar , and the contribution to the charge distribution from this limit should be similar to that of the proton ; a broad distribution of net positive charge . for @xmath127 and above , the @xmath108 quark distribution is less than half the @xmath1 quark distribution , yielding a net negative contribution to the charge . because the distribution of quarks is more localized near @xmath70 as @xmath0 increases , a negative peak can be formed if there is a sufficiently large contribution from down quarks at large @xmath0 values . above @xmath128 , @xmath129 is at least three , except for the density - based extrapolation ( blue circles ) , and continues to increase with @xmath0 . as discussed in ref . @xcite , even if one uses a density - based extrapolation of the emc effect to apply nuclear corrections , the implementation used in this analysis significantly overestimates the effect for deuterium , and yields an unrealistically large result at very large @xmath0 values . in this region , the net impact on the charge distribution will be negative , and will be peaked at smaller values of @xmath69 . this is shown in fig . [ fig : neutron_bins ] , which separately shows the contributions to the neutron charge density from @xmath108 and @xmath1 quarks based on the gpd of ref . the distributions of refs . @xcite yield somewhat different results , but exhibit the same qualitative behavior . for example , the gpds of ref . @xcite , shown in fig . [ fig : neutron_bins ] , yield a negative central neutron charge density for values of @xmath0 between 0.15 and 0.3 and between 0.3 and 0.465 , but for the gpds of ref . @xcite , the central density is positive unless @xmath0 is slightly greater than 0.465 . and @xmath1 quark contributions to the neutron transverse charge density , @xmath130 and @xmath131 . here the quark flavor refers to the neutron ( @xmath108 in the proton is @xmath1 in the neutron ) . the curves correspond to the same @xmath118 regions as in fig . [ fig : proton_scaled ] . the largest contributions come from small @xmath0 , where @xmath108 and @xmath1 quarks contribute similar amounts of charge . as one goes to larger @xmath0 values , the charge is shifted to smaller values of @xmath69 , while at the same time the up quark distribution drops rapidly with respect to the down quarks . reprinted with permission from ref . copyright 2008 by the american physical society.,width=302 ] next we examine the total charge distribution of the neutron . [ fig : neutron_total ] shows the charge distribution of the neutron , separating out the contributions from low and high @xmath0 regions , and shows @xmath132 to suppress the very large density at the center . for @xmath133 the charge distribution ( dotted line ) is positive for @xmath134 fm and slightly negative distribution at larger radii . for @xmath135 , the contribution ( dashed line ) is largely negative , and highly localized below 0.5 fm . the negative region at the center of the neutron transverse charge distribution arises a natural consequence of the model - independent definition of the charge density . the low momentum partons have a larger spatial extent and reproduce the intuitive result of the pion cloud picture : a positive core with a small negative tail at large distances , although the negative tail is difficult to see for this parameterization of the gpd , given the large scale required to show the negative core at small @xmath69 . , dashed is that for @xmath135 , and the solid is the total . reprinted with permission from ref . copyright 2008 by the american physical society.,width=302 ] we also study the quantity _ ( b , x)_qe_q^q_(b , x)to obtain a pictorial view of the transverse charge density for specific values of @xmath0 ( fig . [ pt5 ] ) . the striking feature of the negative spike appears prominently for @xmath127 and more prominently for @xmath128 . these figures show how the central negative charge density appears more and more prominent as @xmath0 increases . clearly the negatively charged @xmath1 quarks dominate at the center of the neutron . for @xmath136 ( top ) , @xmath127 ( middle ) , and @xmath128 ( bottom).,title="fig:",width=302,height=170 ] for @xmath136 ( top ) , @xmath127 ( middle ) , and @xmath128 ( bottom).,title="fig:",width=302 ] for @xmath136 ( top ) , @xmath127 ( middle ) , and @xmath128 ( bottom).,title="fig:",width=302 ] as discussed earlier , the longitudinal - momentum weighting used in determining the nucleon center of mass leads to a strong correlation between the position of the struck quark and the center of the nucleon for large @xmath0 . it is informative to try and remove this effect to obtain something that is closer to our intuitive picture . we can do this by examining the position of the struck quark relative to the center of the _ spectator _ system , so that the struck quark does not influence the definition of the center of the neutron . this can be approximated by looking at the position of the struck quark relative to the spectators . we use eq . ( [ tcm ] ) with the origin set to the center of momentum , for a struck quark at @xmath137 , to determine the momentum - weighted spectator position , @xmath138 , and the relative distance from the struck quark to the spectator quarks : x_1 _ 1+_i>1x_i_i = x+(1-x)_spec=0 , + * b_rel * = - * b_spec*==*b_rel . we exhibit the dependence on @xmath139 by defining a function ^spec_(*b*_rel , x)_(_rel(1-x),x)[redefine ] which gives the probability that a struck quark of longitudinal momentum fraction @xmath0 is a distance @xmath139 away from the spectator center of momentum . figure [ fig : neutron_spectator ] shows this rescaled version of @xmath140 , with the contribution at each @xmath0 value normalized to unity at @xmath70 . the quantity @xmath141 does not correspond to a true density , but can provide a better approximation to our intuitive picture of the charge distribution , as it removes the influence of the struck quark on defining the center of the nucleon . while the charge distribution coming from very low @xmath0 quarks has a greater spatial extent , the decreasing width of the @xmath140 distribution for large @xmath0 quarks is essentially completely removed when looking at @xmath142 . * and @xmath1 quark contributions to @xmath143 ( [ redefine ] ) . vs @xmath142 for @xmath136 ( solid ) , 0.3 ( long - dash ) , 0.5 ( short - dash ) , and 0.7 ( dotted ) . the curves are scaled to unity at @xmath144 . here the quark flavor refers to the neutron ( @xmath108 in the proton is @xmath1 in the neutron ) . reprinted with permission from ref . copyright 2008 by the american physical society.,width=302 ] we summarize our findings with the statement that , using the model gpds of refs . @xcite , the dominance of the neutron s @xmath1 quarks at high values of @xmath0 leads to a negative contribution to the charge density which , due to the definition of @xmath47 , becomes localized near the center of mass of the neutron . this localization does not appear when examined as a function of the position of the struck quark relative to the spectators , and is an consequence of the fact that quarks with a large longitudinal momentum play an important role in defining the transverse position of the neutron . this work was supported by the u. s. department of energy , office of nuclear physics , under contracts fg02 - 97er41014 and de - ac02 - 06ch11357 . we thank d. geesaman , r. holt , p. kroll , c. roberts , m. vanderhaeghen , and b. wojtsekhowski for useful discussions . we thank the ect * for hosting a workshop where many of the calculations we present were performed . h. gao , int . j. mod . e * 12 * ( 2003 ) 1 [ erratum - ibid . e * 12 * ( 2003 ) 567 ] . c. e. hyde - wright and k. de jager , ann . nucl . part . sci . * 54 * , ( 2004 ) 21 . c. f. perdrisat , v. punjabi and m. vanderhaeghen , prog . part . * 59 * ( 2007 ) 694 . j. arrington , c. d. roberts and j. m. zanotti , j. phys . g * 34 * ( 2007 ) s23 g. a. miller , phys . * 99 * ( 2007 ) 112001 . g. a. miller and j. arrington , phys . c * 78 * ( 2008 ) 032201 . soper , phys . d * 5 * ( 1972 ) 1956 ; j. kogut and d.e . soper , phys . rev . d * 1 * ( 1970 ) 2901 . d. e. soper , phys . d * 15 * ( 1977 ) 1141 . m. burkardt , int . j. mod . a * 18 * ( 2003 ) 173 . m. diehl , eur . j. c * 25 * ( 2002)223 , [ erratum - ibid . c * 31 * ( 2003 ) 277 ] . a. w. thomas , s. thberge and g. a. miller , phys . d * 24 * ( 1981 ) 216 ; g. a. miller , phys . c * 66 * ( 2002 ) 032201 . j. l. friar , part . nucl . * 4 * , 153 ( 1972 ) ; r. d. carlitz , s. d. ellis and r. savit , phys . b * 68 * , 443 ( 1977 ) ; n. isgur , g. karl and d. w. l. sprung , phys . d * 23 * , 163 ( 1981 ) . x. d. ji , phys . d * 55 * ( 1997 ) 7114 a. v. radyushkin , phys . d * 56 * ( 1997 ) 5524 m. burkardt , phys . d * 62 * ( 2000 ) 071503 [ erratum - ibid . d * 66 * ( 2002 ) 119903 ] m. guidal , m. v. polyakov , a. v. radyushkin and m. vanderhaeghen , phys . d * 72 * ( 2005 ) 054013 m. diehl , t. feldmann , r. jakob and p. kroll , eur . j. c * 39 * , 1 ( 2005 ) s. ahmad , h. honkanen , s. liuti and s. k. taneja , phys . d * 75 * , 094003 ( 2007 ) b. c. tiburzi , w. detmold and g. a. miller , phys . rev . d * 70 * , 093008 ( 2004 ) j. pumplin , d. r. stump , j. huston , h. l. lai , p. m. nadolsky and w. k. tung , jhep * 0207 * , 012 ( 2002 ) g. a. miller , b. m. k. nefkens and i. slaus , phys . rept . * 194 * , 1 ( 1990 ) . g. a. miller , phys . c * 57 * , 1492 ( 1998 ) j. t. londergan and a. w. thomas , prog . part . phys . * 41 * , 49 ( 1998 ) g. a. miller , a. k. opper and e. j. stephenson , ann . nucl . part . sci . * 56 * , 253 ( 2006 ) l. w. whitlow , e. m. riordan , s. dasu , s. rock and a. bodek , phys . b * 282 * , 475 ( 1992 ) . w. melnitchouk and a. w. thomas , phys . b * 377 * , 11 ( 1996 ) j. arrington , f. coester , r. j. holt and t. s. lee , j. phys . g * 36 * , 025005 ( 2009 )
we find an interpretation of the recent finding that the central charge density of the neutron is negative by using models of generalized parton distributions at zero skewness to relate the behavior of deep inelastic scattering quark distributions , evaluated at high @xmath0 , to the transverse charge density evaluated at small distances . the key physical input of these models is the drell - yan - west relation we find that the @xmath1 quarks dominate the neutron structure function for large values of bjorken @xmath0 , where the large longitudinal momentum of the struck quark has a significant impact on determining the center - of - momentum of the system , and thus the `` center '' of the nucleon in the transverse position plane . # 1#2#3#4#1 * # 2 * , # 3 ( # 4 ) # 1@xmath2#1
band random matrices ( brm ) represent an effective model for both 1d disordered systems with long - range hopping and quasi-1d wires@xcite . the bandwidth @xmath1 plays the role of the range of the interaction in the first case , the one of the square root of the number of independent conduction channels in the second . up to now , studies have been mostly devoted to the analysis of the stationary solutions of the schrdinger equation and to the corresponding spectral properties of brm s @xcite . much less is known about the solutions of the time dependent schrdinger equation , a topic on which only a few studies have been performed @xcite . the partial analogy of this latter problem with the ` dynamical localization ' phenomenon in the kicked rotor @xcite suggests that an initial delta - like packet spreads diffusively and eventually saturates to a localized state . the width of this asymptotic packet for brm s is of the order of @xmath9 lattice sites , i.e. the same order as the localization length of all the eigenfunctions @xcite . the theoretically predicted scaling laws for the mean square displacement @xmath2 were tested numerically and a comparison of the asymptotic form of the wave - packet with a theoretical formula @xcite derived for the 1d anderson model was attempted @xcite . more recently some new theoretical results appeared which give a formula for the time asymptotic packet in the brm model in the large @xmath1 limit @xcite . therefore , it became important to check numerically this formula and to both investigate how the packet reaches its time asymptotic shape and measure the size of finite @xmath1 corrections . for what the time evolution is concerned some phenomenological expressions were suggested in ref . @xcite , based on a power - law convergence of the mean square displacement to its steady state value . however , the presence of large statistical fluctuations prevented the authors of ref . @xcite from assessing whether the time asymptotic scaling is ruled by power law corrections or by the logarithmic corrections to the @xmath10 dependence suggested by rigorous results obtained for the 1d anderson model @xcite . the fluctuations @xmath3 of the width of the asymptotic wave - packet with the realization of the disorder constitute an even more controversial issue , since not even the scaling behaviour is clearly understood . some evidence of an anomalous behaviour was presented in two previous studies of the same problem @xcite and in the kicked rotor @xcite . in all cases the numerics was too poor to make a convincing statement about the value of the anomalous exponent . the bottleneck of the previous simulations was the slowness of the integration scheme , a 4-th order runge - kutta with a small time step to obtain a good conservation of probability over a long time span . this low efficiency prevented from reaching sufficiently large values of @xmath1 and from considering a large enough number of realizations of the brm s . we have instead implemented a 2-nd order cayley algorithm , which , being unitary , exactly conserves probability , although the one - step integration error is larger than the one of the runge - kutta scheme ( a situation similar to those of symplectic algorithms in classical hamiltonian dynamics ) . this has allowed us to more than double the maximum bandwidth ( from @xmath11 to @xmath12 ) and to increase the statistics by a factor four ( in the worst case ) . as a result , we have been able to complete an accurate analysis of the time evolution of the mean square displacement , finding that there is no need to invoke effective formulas with a power - law time dependence even at relatively short times . we have found a clean evidence of an anomalous scaling of the relative fluctuations of the packet width , which behave as @xmath13 with @xmath5 . in order to confirm this anomaly , we have investigated the statistics of the packet width at a specific time in the localization regime . the probability distributions at various @xmath1 values , when appropriately rescaled , superpose , and the resulting universal ( @xmath1 independent ) curve is definitely different from a gaussian with an exponential tail at large @xmath2 values . finally , we have compared our results with the theoretical formula for the asymptotic wave - packet @xcite finding a convincing agreement . the finite @xmath1 corrections to the @xmath7 zhirov expression are of order @xmath14 . we have considered the time - dependent schrdinger equation @xmath15 where @xmath16 is the probability amplitude at site @xmath17 and the tight - binding hamiltonian @xmath18 is a real symmetric band random matrix . the band structure of the hamiltonian is determined by the condition @xmath19 the parameter @xmath1 setting the band - width ; the matrix elements inside the band are independent gaussian random variables with @xmath20 where the symbol @xmath21 stands for the average over different realizations of the disorder . in the present work we have considered the evolution of an ` electron ' initially localized at the centre ( identified with the site @xmath22 ) of an infinite lattice . to this aim we have analysed the solution of equation ( [ schroe ] ) corresponding to the initial condition @xmath23 since the wave - packet evolves in a supposedly infinite lattice , it is necessary to avoid any spurious boundary effect due to the inevitably finite size of the vectors used in the numerical computations . this goal has been achieved by resorting to a self - expanding lattice , i.e. a lattice whose size is progressively enlarged according to the development of the wave - function . at each integration step , our program checks the probability that the electron is in the leftmost and rightmost @xmath1 sites , adding @xmath24 new sites whenever the amplitude @xmath25 is larger than @xmath26 in at least one of the @xmath27 outermost sites . we have separately verified that @xmath28 is small enough not to significantly affect the computation of the probability distribution . for instance , by lowering @xmath28 by an order of magnitude , the mean squared displacement ( computed over the same disorder realizations ) changes only by a few percent . since this systematic error is not larger than the uncertainty due to statistical fluctuations , it is not convenient to reduce the cut - off as it would turn out in a slower code with a consequent reduction of the statistics . the schrdinger equation ( [ schroe ] ) was integrated by approximating the evolution operator @xmath29 with the cayley form @xmath30 which implies that the values of the wave - function at two successive time - steps are related by @xmath31 solving the band diagonal system of equations ( [ cayley2 ] ) allows one to determine @xmath32 once @xmath33 is known . cayley s algorithm is a standard tool for the computation of the solutions of the schrdinger equation ( see for instance ref . @xcite ) ; to the best of our knowledge , this is the first application to the specific field of random hamiltonians with long - range hopping . cayley s form ( [ cayley ] ) for the evolution operator has two relevant features : it is second - order accurate in time and unitary ; in addition , the corresponding integration scheme ( [ cayley2 ] ) is stable . stability is essential in order to study the long time evolution of the wave - packet ; as for unitarity , it ensures the conservation of probability and , together with second - order accuracy in time , allows one to choose time steps @xmath34 two or three order of magnitude bigger than those used in runge - kutta integration schemes . indeed , we could make use of a time step @xmath35 , to be compared with the time step @xmath36 used for the same problem in refs . @xcite . to ascertain how large a @xmath34 could be used , we have compared the solutions obtained through cayley s algorithm at various @xmath34 with the exact solution of the schrdinger equation ( [ schroe ] ) , computed by diagonalizing the hamiltonian ( to avoid boundary effects due to the finite size of the diagonalized matrices , we have considered sufficiently short evolution times ) . by this way we came to the somewhat surprising conclusion that the validity range of the approximate equality ( [ cayley ] ) extended up to time steps as big as @xmath37 ( the scaling of @xmath38 with the band - width @xmath1 was necessary to compensate the opposite scaling of the energy eigenvalues with @xmath39 ) . to check this conclusion , we have computed the mean squared displacement in the localized regime for several values of @xmath34 in the range @xmath40 , finding differences of a few percent , not larger than the statistical fluctuations . this can depend on the fact that the long time evolution of the wave - packet seems to be led by the eigenstates at the band centre . indeed , in the energy representation , the exact evolution operator and the cayley form can be written as @xmath41 with @xmath42 where @xmath43 is the eigenvector corresponding to the energy @xmath44 . these equations show that , for increasing @xmath34 , the approximate equality ( [ cayley ] ) holds true only in the subspace spanned by the eigenvectors corresponding to the band centre , while at the band edge , where the eigenvalues @xmath45 tend to @xmath39 , the coefficients @xmath46 and @xmath47 become quickly different . therefore , the eigenstates at the band edges appear to play a minor role in the time evolution . this is probably due to the shorter localization length of such states compared with those in the centre : in fact , for the mean square displacement , all eigenstates are weighted with their localization length @xcite . to investigate the time evolution of the wave - packet , we have computed the mean square displacement @xmath48 previous studies of this problem strongly suggest that @xmath49 satisfies the scaling relation @xmath50 for large enough values of the bandwidth @xmath1 . nevertheless , in ref . @xcite , where the most detailed numerical investigation has been carried out , it was not possible to obtain a clear verification of the scaling law ( [ scal ] ) due to the poor statistics and the small values of @xmath1 . the faster integration algorithm described in the previous section has allowed us both to average over more realizations ( 400 in the worst case ) , i.e to reduce statistical fluctuations , and to reach larger values of @xmath1 ( namely , @xmath12 instead of @xmath11 as in ref . the results reported in fig . [ fig1 ] for several values of @xmath1 are clean enough to show a convincing convergence from above to a limit shape . in other words , there is no possibility to interpret the deviations as a signature of a different scaling behaviour for @xmath49 . in order to perform a more quantitative analysis , we have proceeded in the following way : @xmath51 to underline the residual but asymptotically irrelevant dependence on the band - with . ] has been averaged over the time interval @xmath52 to obtain the more statistically reliable quantity @xmath53 . by assuming a dependence of the type @xmath54 we have fitted the three parameters @xmath55 , @xmath56 and @xmath57 , finding that the convergence rate @xmath57 is very close to 1 ( 0.95 ) , i.e. that the finite - band corrections are of the order @xmath6 . the fitted value of @xmath55 is 0.61 . the results for the finite - band correction @xmath58 are plotted in fig . [ fig2 ] . the good quality of our numerical data suggests also the possibility to compare the temporal behaviour with the available theoretical formulas . in particular , it has been argued in ref . @xcite that the existence of the so - called mott states should imply a @xmath59 convergence of @xmath60 to its asymptotic value . therefore , we propose the following expression @xmath61 which is the simplest formula that we have found able to reproduce also the initially linear ( i.e. diffusive ) regime . for each value of @xmath1 , the best fit is so close to the numerical data of fig.[fig1 ] to be almost indistinguishable from them ( this is why we do not report the fits on the same figure ) . the meaningfulness of the above expression is further strengthened by the stability of the three free parameters @xmath62 , @xmath63 and @xmath64 , which allows the calculation of a @xmath1-independent diffusion constant ( see below ) . from the values of @xmath62 , we can extrapolate the asymptotic value @xmath65 in exactly the same way as we have done for @xmath66 , finding once more a @xmath14-convergence to a value around 0.70 . this result is to be compared with the theoretical prediction @xmath67 @xcite . the deviation of about 0.03 can be attributed to the accuracy of the integration algorithm . another important parameter that can be extracted from formula ( [ fit2 ] ) is the diffusion coefficient . indeed , by expanding eq . ( [ fit2 ] ) for small @xmath68 , we find @xmath69 the diffusion constant @xmath70 turns out to be close to 0.50 for all values of @xmath1 and , what is more important , close to the value that we obtain from a quadratic fit of the initial growth rate of the packet . this is a very encouraging result , since it confirms the correctness of formula ( [ fit2 ] ) for both the diffusive and the localized regimes . let us notice that the value @xmath71 is somewhat smaller than the one reported in @xcite ( @xmath72 ) . taking into account statistical fluctuations and systematic deviations , we find that @xmath73 . in past papers , a phenomenological expression involving a power - law convergence to the asymptotic value of the mean squared displacement has been proposed @xcite , arguing that it should provide an effective description of both the diffusive and localized regime @xmath74 the success of expression ( [ fit2 ] ) shows that there is no need to introduce anomalous power laws to reproduce the numerical findings . however , for the sake of completeness , we have fitted our numerical data also with eq . ( [ fit3 ] ) , finding an equally good agreement . therefore , on the basis of the quality of the fit we can not conclude which of the two expressions is better ; nevertheless , it is worth recalling that the former one has the correct asymptotic behaviour and , moreover , the fitted parameters are more stable . a much more controversial situation exists about the fluctuations of the packet width . let us introduce the r.m.s . deviation @xmath75 in fact , it has not yet been clarified how the above variable scales in the large-@xmath1 limit . in particular , the correct value of the scaling exponent @xmath76 in the relation @xmath77 is still unknown ; this is why the @xmath1 dependence in @xmath78 is explicitly maintained . a scaling like @xmath79 would imply that the packet - width is not a self - averaging quantity , since the relative size of the fluctuations would not go to zero for increasing @xmath1 . conversely , an exponent @xmath80 corresponds both to self - averaging and ` normal ' behaviour . in fact , the number @xmath81 of independent channels ( lattice sites ) actively contributing to the localized region ( i.e. the localization length ) is of the order @xmath9 . if we assume that all such contributions to the second moment @xmath2 are independent of one another , then we are led to conclude that the relative fluctuations should decrease as @xmath82 , thus yielding an absolute growth as @xmath83 . since previous studies @xcite have suggested a small anomaly , i.e. @xmath76 slightly larger than 3 , we have chosen to report the behaviour of @xmath84 for @xmath80 . the data shown in fig . [ fig3 ] reveals a drastically different behaviour from what observed in fig . first of all , the curves tend to grow for increasing @xmath1 ; moreover , there is no obvious indication of a convergence to some finite value . altogether , these features imply that @xmath76 is strictly larger than 3 , qualitatively in agreement with previous simulations . in order to perform a more quantitative analysis , we have computed the average of @xmath3 and rescaled it to the average of @xmath49 , @xmath85 ( @xmath86 is again to be interpreted as the average over the time interval @xmath52 ) . the advantage of this renormalization , already adopted in ref . @xcite , is that it reduces finite - band corrections . the results reported in fig . [ fig4 ] , reveal a clean power law decay with an exponent @xmath87 . this value is slightly larger than the one found in the previous studies , but follows from a much cleaner numerics . a more global check of the scaling behaviour can be made by plotting the rescaled fluctuations @xmath88 for the various values of @xmath1 . the optimal value of @xmath0 can thus be identified as that one yielding the best data collapse . the curves reported in the inset of fig . [ fig4 ] have been obtained for @xmath89 . it is necessary to modify @xmath0 by at least @xmath90 units in order to see a significant worsening of the data collapse . accordingly , the best estimate of the anomalous exponent is @xmath5 , so that the dependence of @xmath78 on @xmath1 in eq . ( [ scal2 ] ) is removed for @xmath91 . in order to find further support for this anomalous behaviour , we have investigated the probability distribution @xmath92 for the second moment at the time @xmath93 ( the longest time we reached for the larger @xmath1-values ) , i.e. when the wave - function has almost entered the steady - state regime . the construction of reliable histograms has forced us to consider smaller values of @xmath1 . in fact , we have studied the cases @xmath94 , @xmath11 and @xmath95 , using @xmath96 realizations of the disorder in the first two cases and @xmath97 in the last one . the results are reported in fig . [ fig5 ] , where , following a method suggested in ref . @xcite , we have conveniently rescaled the probability distribution @xmath92 . in particular , defining by @xmath98 the average value of @xmath60 at @xmath99 and @xmath100 the standard deviation over the ensemble of disorder realizations , we have plotted @xmath101 vs. @xmath102 . after this rescaling , the distributions @xmath103 , corresponding to the three @xmath1 values , have zero average and unit standard deviation . it is remarkable to notice that all curves nicely overlap indicating a striking scaling behaviour . a further important feature is the deviation from a gaussian behaviour , especially for large values of @xmath104 , where a clear exponential tail is visible . the dotted line just above the three curves ( corresponding to the pure exponential @xmath105 ) has been added to give an idea of the decay rate which is slightly larger than 1 . the results of this analysis are important in two respects : i ) the exponential tail ` explains ' the difficulties encountered in getting rid of statistical fluctuations in the estimate of @xmath78 ; ii ) the deviations from a gaussian behaviour provide an independent evidence of the anomalous scaling behaviour of the fluctuations . it is interesting to remark that a preliminary quantitative comparison has revealed a striking identity of the probability @xmath103 with the distributions found in ref . @xcite for such diverse quantities as the magnetization in the 2d xy model and the power consumption in a confined turbulent flow . this close correspondence deserves further investigations . finally , we want to compare our results with the theoretical predictions for the asymptotic shape of the wave - packet . in @xcite a reasonable agreement was found between the numerical data and the formula obtained by gogolin for strictly one - dimensional systems @xcite . since a theoretical expression has been derived in the meantime for quasi-1d systems @xcite , it is desirable to compare our data also with this expression . in fig . [ fig6 ] we present the disorder - averaged probability profiles @xmath106 for large times , rescaled under the assumption @xcite @xmath107 and compare them with zhirov s theoretical formula , which is denoted by the white line . no appreciable deviation is noticeable except for the extreme part of the tails , where it is reasonable to expect numerical errors due to boundary effects . the good overlap is partly due to the ( unavoidable ) choice of logarithmic scales in fig . . however , if we zoom the region around the maximum ( with the exception of the zero channel ) , one can see in fig . [ fig7 ] a slow tendency of the various curves to grow towards the theoretical expectation . this is consistent with the behaviour of @xmath60 reported in fig . [ fig1 ] , which reveals a convergence from above for the mean squared displacement . it is interesting to notice that all such deviations are mainly due to the finiteness of @xmath1 while the lack of asymptoticity in @xmath108 appears to be much less relevant . finally , we consider separately the zero channel , i.e. the return probability to the origin @xmath109 . in fig . [ fig8 ] we plot this quantity versus @xmath68 for different values of @xmath1 . in all cases , a quite fast convergence , as compared with the behaviour of the packet - width , to the asymptotic value is clearly seen . in practice , as soon as @xmath68 is about 1 , the average value of @xmath109 reaches the asymptotic value . it is instructive to compare our numerical findings with the asymptotic ( in time and @xmath1 ) analytic expression @xmath110 . by fitting the dependence of the time average of @xmath111 ( in the interval @xmath112 ) on @xmath1 as in eq . [ fit1 ] ) , we find that the asymptotic value is about 5.7 and that the convergence rate is @xmath113 . the numerical value is in a reasonable agreement with the theoretical one , considering that it is the result of an extrapolation of data already affected by errors of the order of a few percent . the non trivial part of the result is the rate of convergence of this probability , which is definitely slower than the @xmath6 behaviour displayed by the second ( and other low order ) moments . the behaviour of the return probability , however , is in agreement with the theoretical predictions made in @xcite , where the finite @xmath1 deviations from the asymptotic steady - state probability distribution were estimated to be of order @xmath8 in the @xmath114 neighbourhood of the origin . the @xmath1-dependence of the return probability is further illustrated in fig . [ fig9 ] , where we plot the deviation @xmath115 from the asymptotic value @xmath116 as a function of @xmath1 in bilogarithmic scale . the displayed numerical values were obtained by averaging the return probability both over disorder realizations and the time interval @xmath117 ; the data were then fitted with two expressions , exhibiting deviations from the asymptotic value @xmath118 of order @xmath119 and @xmath120 respectively . in the second case , the exponent @xmath57 was used as a fitting parameter and the best fit value was @xmath121 . as can be seen from fig . [ fig9 ] , the power law with @xmath8 corrections fits the data much better than the one with deviations of order @xmath119 . we have studied the time evolution of an initial @xmath0-like wave - packet in a 1d disordered lattice with long - range hopping . the main results of this paper are the following . we have confirmed with clean numerics the scaling law ( [ scal ] ) for the mean square displacement @xmath49 , first proposed and studied in ref . this scaling law is valid in the large @xmath1 limit ; here we have found that finite @xmath1 corrections are of the order @xmath6 . we have proposed formula ( [ fit2 ] ) for fitting the time evolution of @xmath49 towards its steady state value ; this formula contains the logarithmic corrections suggested by the existence of mott states . we confirm the presence of an anomaly in the scaling law of the relative fluctuations @xmath122 of the mean square displacement , finding that they vanish for large @xmath1 as @xmath123 . we have linked this anomaly to the presence of non - gaussian fluctuations of the mean square displacement . in fact , the probability distribution of @xmath49 displays an exponential tail for large values of @xmath49 . the conveniently rescaled probability strikingly coincides with the distributions obtained in ref . @xcite for such diverse quantities as the magnetization in the 2d xy model and the power consumption in a confined turbulent flow . the degree of universality of such distribution deserves further investigations . finally , we have compared the numerical results on the steady state probability profile with the theoretical formula proposed by zhirov for large @xmath1 , finding a good agreement . we have computed for the first time finite @xmath1 corrections , obtaining @xmath119 deviations for the moments of the probability profile and @xmath8 corrections for the return probability to the origin . we thank b. chirikov for having suggested the numerical verification of zhirov s formula . we thank o. zhirov for having exchanged with us his ideas and for his comments on our numerical results . we finally thank f.m . izrailev , h. kantz and b. mehlig for useful discussions . t. guhr , a. mller - groeling and h.a . weidenmller , _ phys . _ , * 299 * , 189 ( 1998 ) and refs . therein . g. casati , l. molinari and f.m . izrailev , _ phys . _ , * 64 * , 1851 ( 1990 ) ; g. casati , f.m . izrailev and l. molinari , _ a _ , * 24 * , 4755 ( 1991 ) ; y. v. fyodorov and a.d . mirlin , _ phys . _ , * 67 * , 2405 ( 1991 ) . f. m. izrailev , t. kottos , a. politi , s. ruffo , g. p. tsironis , _ europhys . _ , * 34 * , 441 , ( 1996 ) . f. m. izrailev , t. kottos , a. politi , g. p. tsironis , _ phys . e _ , * 55 * , 4951 , ( 1997 ) . g. casati , i. guarneri , f.m . izrailev and r. scharf , _ phys . _ , * 64 * , 5 ( 1990 ) and refs . therein gogolin , zh . * 71 * , 1912 ( 1976 ) [ sov . jetp , * 44 * , 1003 ( 1976 ) ] ; _ phys . _ , * 86 * , 1 ( 1981 ) . o. v. zhirov , _ chao - dyn 9712001_. g. casati and b. chirikov , in _ quantum chaos : between order and disorder _ , g. casati and b. chirikov eds . , cambridge university press , cambridge ( 1994 ) . e.p.nachmedov , v.n.prigodin and u.a.firsov , zh . * 92 * 2133 ( 1987 ) [ sov . jetp , * 65 * 1202 ( 1987 ) ] . w. press et al . , _ numerical recipes in fortran : the art of scientific computing _ , second edition , cambridge university press , cambridge , ( 1992 ) . we are indebted to an anonymous referee for this observation . f. m. izrailev , _ phys . rep . _ , * 196 * , 299 , ( 1990 ) o. v. zhirov , _ private communication_. s.t . bramwell , p.c . holdsworth and j.f . pinton , _ nature _ , * 396 * , 552 ( 1998 ) .
we have investigated numerically the quantum evolution of a @xmath0-like wave - packet in a quenched disordered medium described by a tight - binding hamiltonian with long - range hopping ( band random matrix approach ) . we have obtained clean data for the scaling properties in time and in the bandwidth @xmath1 of the packet width @xmath2 and its fluctuations @xmath3 with respect to disorder realizations . we confirm that the fluctuations of the packet width in the steady - state show an anomalous scaling @xmath4 with @xmath5 . this can be related to the presence of non - gaussian tails in the distribution of @xmath2 . finally , we have analysed the steady state probability profile and we have found @xmath6 corrections with respect to the theoretical formula derived by zhirov in the @xmath7 limit , except at the origin , where the corrections are @xmath8 . pacs numbers : 05.45.mt , 71.23.an , 72.15.rn , 05.60.gg
density functional theory ( dft ) , in both its static @xcite and time - dependent @xcite forms , has become by far the most successful and widely used among all the electronic structure methods . the most obvious reason for this success is that it provides accurate predictions of numerous properties of atoms , inorganic molecules , bio - molecules , nanostructures and solids , thus serving different scientific communities . in addition dft has a solid theoretical foundation . the hohenberg - kohn theorem @xcite establishes the existence of a unique energy functional @xmath1 $ ] of the electron density @xmath2 which alone is sufficient to determine the exact ground - state of a @xmath3-electron system . although the energy functional itself is not known , several of its general properties can be demonstrated rigorously . these are crucial for constructing increasingly more predictive approximations to the functional and for addressing the failures of such approximations@xcite . finally , but no less important , the kohn - sham ( ks ) formulation of dft @xcite establishes a one to one mapping of the intrinsically many - body problem onto a fictitious single - particle system and offers a convenient way for minimizing @xmath1 $ ] . the degree of complexity of the kohn - sham ( ks ) problem depends on the approximation chosen for the density functional . in the case of the local density approximation ( lda ) @xcite the ks problem typically scales as @xmath4 , where the scaling is dominated by the diagonalization algorithm . however , clever choices with regards to basis sets and sophisticated numerical methods make order-@xmath3 scaling a reality @xcite . the energy functional @xmath5 $ ] ( @xmath6 , @xmath7 is the spin density , @xmath8 ) can be written as @xmath9=t_\mathrm{s}[\rho]+\int\mathrm{d}^3{\bf r}\:\rho({\bf r})v({\bf r})+u[\rho]+e_\mathrm{xc}[\rho^\uparrow,\rho^\downarrow]\ ; , \label{enfunc}\ ] ] where @xmath10 is the kinetic energy of the non - interacting system , @xmath11 the external potential , @xmath0 the hartree electrostatic energy and @xmath12 the exchange and correlation ( xc ) energy . this last term is unknown and must be approximated . the construction of an approximated functional follows two strategies : empirical and `` constraint satisfaction '' . empirical xc functionals usually violate some of the constraints imposed by exact dft , and rely on parameterizations obtained by fitting representative data . one includes in this category , xc functionals which borrow some functional dependence from other theories . this is for instance the case of the celebrated lda+@xmath0 scheme @xcite , where the hubbard-@xmath0 energy takes the place of the lda energy for certain `` strongly correlated '' atomic orbitals ( typically @xmath13 and @xmath14 shells ) . the method however depends on the knowledge of the coulomb and exchange parameters @xmath0 and @xmath15 , which vary from material to material , and can also be different for the same ion in different chemical environments @xcite . in contrast the construction based on `` constraint satisfaction '' proceeds by developing increasingly more sophisticated functionals , which nevertheless satisfy most of the properties of exact dft @xcite . it was argued that this method constructs a `` jacob s ladder '' @xcite , where functionals are assigned to different rungs depending on the number of ingredients they include . thus the lda , which depends only on the spin - densities is on the first rung , the generalized gradient approximation ( gga ) @xcite , which depends also on @xmath16 , is on the second rung , the so - called meta - gga functionals @xcite , which in addition to @xmath6 and @xmath16 depend on either the laplacian @xmath17 or the orbital kinetic energy density , are on the third rung and so on . the higher its position on the ladder the more accurate a functional becomes , but at the price of increasing computational overheads . therefore its worth investigating corrections to the functionals of the lower rungs , which preserve most of the fundamental properties of dft and do not generate massive additional numerical overheads . one of the fundamental problems intrinsic to the semi - local functionals of the first three rungs is the presence of self - interaction ( si ) @xcite . this is the spurious interaction of an electron in a given ks orbital with the hartree and xc potential generated by itself . such an interaction is not present in the hartree - fock method , where the coulomb self - interaction of occupied orbitals is cancelled exactly by the non - local exchange . however when using semi - local functionals such a cancellation is not complete and the rigorous condition for ks - dft @xmath18+e_\mathrm{xc}[\rho_n^\sigma,0]=0\ ] ] for the orbital density @xmath19 of the fully occupied ks orbital @xmath20 is not satisfied . a direct consequence of the self - interaction in lda / gga is that the ks potential becomes too repulsive and exhibits an incorrect asymptotic behavior @xcite . this `` schizophrenic '' ( self - interacting ) nature of semi - local ks potentials generates a number of failures in describing elementary properties of atoms , molecules and solids . negatively charged ions ( h@xmath21 , o@xmath21 , f@xmath21 ) and molecules are predicted to be unstable by lda @xcite and transition metal oxides are described as small - gap mott - hubbard antiferromagnets ( mno , nio ) @xcite or even as ferromagnetic metals ( feo , coo ) @xcite instead of charge - transfer insulators . moreover the ks highest occupied molecular orbital ( homo ) , the only ks eigenvalue that can be rigorously associated to a single particle energy @xcite , is usually nowhere near the actual ionization potential @xcite . finally xc functionals affected by si do not present a derivative discontinuity as a function of the occupation @xcite . semi - local functionals in fact continuously connect the orbital levels of systems of different integer occupation . this means for instance that when adding an extra electron to an open shell @xmath3-electron system the ks potential does not jump discontinuously by @xmath22 where @xmath23 and @xmath24 are respectively the ionization potential and the electron affinity for the @xmath3-electron system . this serious drawback is responsible for the incorrect dissociation of heteronuclear molecules into charged ions @xcite and for the metallic conductance of insulating molecules @xcite . the problem of removing the si from a semi - local potential was acknowledged a long time ago when fermi and amaldi proposed a first crude correction @xcite . however the modern theory of self - interaction corrections ( sic ) in dft is due to the original work of perdew and zunger from almost three decades ago @xcite . their idea consists in removing directly the self - hartree and self - xc energy of all the occupied ks orbitals from the lda xc functional ( the same argument is valid for other semi - local functionals ) , thus defining the sic functional as @xmath25=e_\mathrm{xc}^\mathrm{lda}[\rho^\uparrow,\rho^\downarrow]- \sum_{n\sigma}^\mathrm{occupied}(u[\rho_n^\sigma]+e_\mathrm{xc}^\mathrm{lda}[\rho_n^\sigma,0])\;. \label{efsic}\ ] ] although apparently simple , the sic method is more involved than standard ks dft . the theory is still a density functional one , i.e. it satisfies the hohenberg - kohn theorem , however it does not fit into the kohn - sham scheme , since the one - particle potential is orbital - dependent . this means that one can not define a kinetic energy functional independently from the choice of @xmath12 @xcite . two immediate consequences are that the @xmath20 are not orthogonal , and that the orbital - dependent potential can break the symmetry of the system . this last aspect is particularly important for solids since one has to give up the bloch representation . in this paper we explore an approximate method for sic to the lda , which has the benefit of preserving the local nature of the lda potential , and therefore maintains all of the system s symmetries . we have followed in the footsteps of filippetti and spaldin @xcite , who extended the original idea of vogel and co - workers @xcite of considering only the atomic contributions to the sic . we have implemented such a scheme into the localized atomic orbital code _ siesta _ @xcite and applied it to a vast range of molecules and solids . in particular we have investigated in detail how the scheme performs as a single - particle theory , and how the sic should be rescaled in different chemical environments . the direct subtraction proposed by perdew and zunger is the foundation of the modern sic method . however the minimization of the sic functional ( [ efsic ] ) is not trivial , in particular for extended systems . the main problem is that @xmath12 itself depends on the ks orbitals . thus it does not fit into the standard ks scheme and a more complicated minimization procedure is needed . following the minimization strategy proposed by levy @xcite , which prescribes to minimize the functional first with respect to the ks orbitals @xmath20 and then with respect to the occupation numbers @xmath26 , perdew and zunger derived a set of single - particle equations @xmath27\psi_n^\sigma=\epsilon_{n}^{\sigma,\mathrm{sic } } \psi_n^\sigma\ ; , \label{kssic}\ ] ] where the effective single - particle potential @xmath28 is defined as @xmath29 ; { \bf r})+v_\mathrm{xc}^{\sigma,\mathrm{lda}}([\rho^\uparrow,\rho^\downarrow ] ; { \bf r})-\nonumber \\ -u([\rho_n ] ; { \bf r})-v_\mathrm{xc}^{\sigma,\mathrm{lda}}([\rho_n^\uparrow,0 ] ; { \bf r } ) \ ; , \label{kssic - pot}\end{aligned}\ ] ] and @xmath30 ; { \bf r})=\int\mathrm{d}^3{\bf r}^\prime\frac{\rho({\bf r}^\prime)}{|{\bf r}-{\bf r}^\prime|}\ ; , \\ v_\mathrm{xc}^{\sigma,\mathrm{lda}}([\rho^\uparrow,\rho^\downarrow ] ; { \bf r})=\frac{\delta}{\delta \rho^\sigma({\bf r } ) } e_\mathrm{xc}^\mathrm{lda}[\rho^\uparrow,\rho^\downarrow]\;.\end{aligned}\ ] ] these are solved in the standard ks way for atoms , with good results in terms of quasi - particle energies @xcite . in this particular case the ks orbitals @xmath20 show only small deviations from orthogonality , which is re - imposed with a standard schmidt orthogonalization . the problem of the non - orthogonality of the ks orbitals can be easily solved by imposing the orthogonality condition as a constraint to the energy functional , thus obtaining the following single - particle equation @xmath27\psi_n^\sigma = \sum_m\epsilon_{nm}^{\sigma,\mathrm{sic}}\psi_m^\sigma\;. \label{kssic - orth}\ ] ] even in this case where orthogonality is imposed , two major problems remain : the orbitals minimizing the energy functional are not ks - type and in general do not satisfy the system s symmetries . if one insists in minimizing the energy functional in a ks fashion by constructing the orbitals according to the symmetries of the system , the theory will become size - inconsistent , or in other words it will be dependent on the particular representation employed . thus one might arrive at a paradox , where in the self - interaction of @xmath3 hydrogen atoms arranged on a regular lattice of large lattice spacing ( in such a way that there is no interaction between the atoms ) vanishes , since the sic of a bloch state vanishes for @xmath31 . however the sic for an individual h atom , when calculated using atomic - like orbitals , accounts for essentially all the ground - state energy error @xcite . therefore a size - consistent theory of sic dft must look for a scheme where a unitary transformation of the occupied orbitals , which minimizes the sic energy is performed . this idea is at the foundation of all modern implementations of sic . significant progress towards the construction of a size - consistent sic theory was made by pederson , heaton and lin , who introduced two sets of orbitals : localized orbitals @xmath32 minimizing @xmath33 and canonical ( kohn - sham ) de - localized orbitals @xmath20 @xcite . the localized orbitals are used for defining the densities entering into the effective potential ( [ kssic - pot ] ) , while the canonical orbitals are used for extracting the lagrangian multipliers @xmath34 , which are then associated to the ks eigenvalues . the two sets are related by unitary transformation @xmath35 , and one has two possible strategies for minimizing the total energy . the first consists in a direct minimization with respect to the localized orbitals @xmath32 , i.e. in solving equation ( [ kssic - orth ] ) when we replace @xmath36 with @xmath37 and the orbital densities entering the definition of the one - particle potential ( [ kssic - pot ] ) are simply @xmath38 . in addition the following minimization condition must be satisfied @xmath39 where @xmath40 ; { \bf r})+v_\mathrm{xc}^{\sigma,\mathrm{lda}}([\rho_n^\uparrow,0 ] ; { \bf r})$ ] . an expression for the gradient of the sic functional , which also constrains the orbitals to be orthogonal has been derived @xcite and applied to atoms and molecules with a mixture of successes and bad failures @xcite . the second strategy uses the canonical orbitals @xmath36 and seeks the minimization of the sic energy by varying both the orbitals @xmath36 and the unitary transformation @xmath41 . the corresponding set of equations is @xmath42 @xmath43 @xmath44 where @xmath45 is the standard lda hamiltonian ( without sic ) . thus the sic potential for the canonical orbitals appears as a weighted average of the sic potential for the localized orbitals . the solutions of the set of equations ( [ kssic - pro ] ) is somehow more appealing than that associated to the localized orbitals since the canonical orbitals can be constructed in a way that preserves the system s symmetries ( for instance translational invariance ) . a convenient way for solving the equation ( [ kssic - pro ] ) is that of using the so called `` unified hamiltonian '' method @xcite . this is defined as ( we drop the spin index @xmath46 ) @xmath47 where @xmath48 is the projector over the occupied orbital @xmath20 , and @xmath49 is the projector over the unoccupied ones @xmath50 . the crucial point is that the diagonal elements of the matrix @xmath34 and their corresponding orbitals @xmath20 are respectively eigenvalues and eigenvectors of @xmath51 , from which the whole @xmath34 can be constructed . finally , and perhaps most important , at the minimum of the sic functional , the canonical orbitals diagonalize the matrix @xmath34 , whose eigenvalues can now be interpreted as an analogue of the kohn - sham eigenvalues @xcite . it is also interesting to note that an alternative way for obtaining orbital energies is that of constructing an effective si - free local potential using the krieger - li - iafrate method @xcite . this has been recently explored by several groups @xcite when applied to extended systems the sic method demands considerable additional computational overheads over standard lda . thus for a long time it has not encountered the favor of the general solid state community . in the case of solids the price to pay for not using canonical orbitals is enormous since the bloch representation should be abandoned and in principle infinite cells should be considered . for this reason the second minimization scheme , in which the canonical orbitals are in a bloch form , is more suitable . in this case for each @xmath52-vector one can derive an equation identical to equation ( [ kssic - pro ] ) , where @xmath53 and @xmath54 is simply the band index @xcite . the associated localized orbitals @xmath37 for instance can be constructed as wannier functions and the minimization scheme proceeds in a similar way to that done for molecules . the problem here is that in practice , the cell needed to describe the localized states @xmath37 may be considerably larger than the primitive unit cell . this is not the case for ionic insulators @xcite , where the localized orbitals are well approximated by atomic orbitals . such a simplification is however not valid in general . for example supercells as large as 500 atoms have been considered for describing the localized @xmath13 shells in transition metals oxides @xcite . despite these difficulties the sic scheme has been applied to a vast range of solid state systems with systematic improvement over lda . these include , in addition to transition metals monoxides @xcite , rare - earth materials @xcite , diluted magnetic semiconductors @xcite , fe@xmath55o@xmath56 @xcite , heavy elements compounds @xcite , just to name a few . in order to make the sic method more scalable several approximations have been proposed . one possibility is that of incorporating part of the sic into the definition of the pseudopotentials @xcite . the idea consists in subtracting the atomic si from the free atom , and then transferring the resulting electronic structure to the definition of a standard norm - conserving pseudopotential . this approximation is sustained by the fact that the transformation @xmath41 , which relates canonical and localized orbitals does hardly mix core and valence states @xcite . thus the sic contribution to the total energy can be separated into the contributions from the core and the valence and in first approximation the latter can be neglected @xcite . the benefit of this method is that translational invariance is regained and the complicated procedure of minimizing @xmath41 is replaced by a pseudopotential calculation . a further improvement over the pseudopotential method was recently presented by vogel and co - workers @xcite and then extended by filippetti and spaldin @xcite . the method still assumes separability between the core and the valence contributions to the sic , but now the sic for the valence electrons is approximated by an atomic - like contribution , instead of being neglected . this atomic sic ( asic ) scheme is certainly a drastic approximation , since it implicitly assumes that the transformation @xmath41 minimizing the sic functional leads to atomic like orbitals . in the work of vogel this additional contribution is not evaluated self - consistently for the solid , while the implementation of filippetti assumes a linear dependance of the sic over the orbital occupation . in spite of the approximations involved , the method has been applied successfully to a range of solids including ii - vi semiconductors and nitrites @xcite , transition metal monoxides @xcite , silver halides @xcite , noble metal oxides @xcite , ferroelectric materials @xcite , high - k materials @xcite and diluted magnetic semiconductors @xcite . interestingly most of the systems addressed by the asic method are characterized by semi - core @xmath13 orbitals , for which an atomic correction looks appropriate , and a similar argument is probably valid for ionic compounds as recently demonstrated for the case of sic @xcite . here we further investigate the self - consistent asic method @xcite by examining both finite and extended systems , and by critically considering whether a scaling factor , additional to the orbital occupation , is needed for reproducing the correct single particle spectrum . in this section we derive the fundamental equations of the asic method , while looking closely at the main approximations involved in comparison to the fully self - consistent sic approach . our practical implementation is also described . the starting point of our analysis is the sic schrdinger - like equation ( [ kssic - pro ] ) for the canonical orbitals . let us assume , as from reference @xcite , that the rotation @xmath41 transforming localized orbitals ( to be determined ) into canonical orbitals ( see equation ( [ sic - rot ] ) ) does not mix core and valence states . we also assume that core electrons are well localized into atomic - like wave - functions and that they can be effectively described by a norm - conserving pseudopotential . let us now assume that @xmath41 is known and so are the localized orbitals @xmath57 . in this case the canonical orbitals diagonalize @xmath34 and the equation ( [ kssic - pro ] ) simply reduces to @xmath58 with @xmath59 defined in equation ( [ dsic ] ) . the hamiltonian @xmath60 can be then re - written in a convenient form as @xmath61 where @xmath62 is the self - interaction potential for the localized orbital @xmath57 , and @xmath63 is the projector over the same state @xmath64 two main approximations are then taken in the asic approach @xcite . first the localized states @xmath57 are assumed to be atomic - like orbitals @xmath65 ( asic orbitals ) and the sic potential is approximated as @xmath66 with @xmath67 ; { \bf r})+v_\mathrm{xc}^{\sigma,\mathrm{lda}}([\rho_m^\uparrow,0 ] ; { \bf r})$ ] and @xmath68 , @xmath69 is the projector of equation ( [ phiproj ] ) obtained by replacing the @xmath37 s with the asic orbitals @xmath70 , and @xmath71 is a scaling factor . note that the orbitals @xmath72 are not explicitly spin - dependent and one simply has @xmath65=@xmath73 with @xmath74 the orbital occupation ( @xmath75 ) . the factor @xmath71 is an empirical factor , which accounts for the particular choice of asic orbitals . this first approximation is expected to be accurate for systems retaining an atomic - like charge density as in the case of small molecules . it is also formally exact in the one - electron limit ( for @xmath76 ) . in the case of extended solids the situation is less transparent , since in general the functions minimizing @xmath33 are wannier - like functions @xcite . the second approximation taken in the asic method is that of replacing the non - local projector @xmath69 with its expectation value . the idea is that the sic potential for the canonical orbitals @xmath59 is formally a weighted average of the sic potential for the localized orbitals @xmath77 . for the exact sic method the weighting factor is the non - local projector @xmath78 . this means that the sic potential for a given canonical orbital @xmath79 is maximized in those regions where the overlap with some of the localized orbitals @xmath80 is maximum . in the asic method such non - local projector is replaced more conveniently by a scalar . in the original proposal by voegl et al . @xcite this was simply set to one . here we consider the orbital occupation @xmath74 of the given asic orbital @xmath72 , i.e. we replace @xmath69 with its expectation value @xmath81 where @xmath82 is the occupation number of the kohn - sham orbital @xmath20 . the final form of asic potential is then @xmath83 let us now comment on the empirical scaling factor @xmath71 . in reference @xcite @xmath71 was set to @xmath84 in order to capture eigenvalue relaxation . this choice however violates the one - electron limit of the sic potential , which is correctly reproduced for @xmath76 . we can then interpret @xmath71 as a measure of the deviation of the asic potential from the exact sic potential . ultimately @xmath71 reflects the deviation of the actual asic projectors @xmath85 from the localized orbitals defining the si corrected ground state . one then expects @xmath71 to be close to 1 for systems with an atomic - like charge density , and to vanish for metals , whose valence charge density resembles that of a uniform electron gas @xcite . intermediate values are then expected for situations different from these two extremes , and we will show that a values around @xmath84 describe well a vast class of mid- to wide - gap insulators . the final form of the sic potential to subtract from the lsda ( local spin density approximation ) one ( equation ( [ psic - appr - final ] ) ) is that of a linear combination of non - local pseudopotential - like terms . these are uniquely defined by the choice of exchange and correlation potential used and by asic orbitals @xmath72 . the practical way of constructing such potentials , i.e. the way of importing the atomic sic to the solid state , depends on the specific numerical implementation used for the dft algorithm . at present only plane - wave implementations are available @xcite , while here we present our new scheme based on the pseudo atomic orbital ( pao ) @xcite code _ siesta _ @xcite . we start by solving the atomic all - electron sic - lsda equation for all the species involved in the solid state calculation . here we apply the original perdew - zunger ( pz - sic ) formalism @xcite and we neglect the small non - orthogonality between the kohn - sham orbitals . thus we obtain a set of si corrected atomic orbitals @xmath72 , which exactly solve the atomic sic - lsda problem . the atomic orbitals @xmath72 describing the valence electrons are then used to define the asic potentials @xmath86 @xmath87 ; { \bf r})+v_\mathrm{xc}^{\sigma,\mathrm{lda}}([\rho_m^\uparrow,0 ] ; { \bf r})\ ] ] with @xmath88 . at the same time a standard lsda calculation for the same atoms is used to construct the pseudopotentials describing the core electrons . these are standard norm - conserving scalar relativistic troullier - martins pseudopotentials @xcite with nonlinear core corrections @xcite . thus we usually neglect the sic over the core states , when constructing the pseudopotentials . this is justified by the fact that the eigenvalues for the sic - lsda - pseudoatom , i.e. for the free atom where the effects of core electrons are replaced by lsda pseudopotentials but sic is applied to the valence electrons , are in excellent agreement with those obtained by all - electron sic - lsda calculations @xcite . the final step is that of recasting the asic potentials @xmath89 , which have a @xmath90 asymptotic behaviour , in a suitable non - local form . this is obtained with the standard kleinman - bylander @xcite scheme and the final asic potential ( equation ( [ psic - appr - final ] ) ) is written as @xmath91 where the asic projectors are given by @xmath92 and the normalization factors are @xmath93 the orbital functions @xmath94 are atomic - like functions with a finite range , which ensure that the asic projectors @xmath95 do not extend beyond that range . these are constructed in the same way as the _ siesta _ basis set orbitals , i.e. as solutions of the pseudo - atomic problem with an additional confining potential at the cutoff radius @xmath96 @xcite . the choice of the appropriate cutoff radius for the sic projectors should take into account the two following requirements . on the one hand it should be sufficiently large to capture most of the sic corrections . a good criterion @xcite is that the sic - lsda contribution to the orbital energy of the free atom @xmath97 is reproduced within some tolerance . on the other hand the cutoff should be reasonably short so as not to change the connectivity of the matrix elements of the pao hamiltonian . in other words we need to ensure that orbitals otherwise considered as disconnected in evaluating the various parts of the hamiltonian matrix are not considered connected for the @xmath98 matrix elements alone . as a practical rule we set the cutoff radius for a particular orbital of a given atom to be either equal to the longest among the cutoff radii of the pao basis set for that particular atom ( typically the first @xmath99 of the lowest angular momentum ) , or , if shorter , the radius at which @xmath100 0.1mry . typically , when reasonable cutoff radii ( 6 to 9 bohr ) are used , we find that the atomic sic - lsda eigenvalues are reproduced to within 1 to 5 mry for the most extended shells and to within 0.1 mry for more confined shells . thus @xmath101 are rather well converged already for cutoff radii defined by a pao energies shifts @xcite of around 20mry , although usually smaller pao energy shifts are necessary for highly converged total energy calculations . finally the matrix elements of the sic potential are calculated as usual over the _ siesta _ basis set . additional basis functions @xmath102 are constructed from the confined localized atomic orbitals described before using the split - norm scheme @xcite . the density matrix @xmath6 is represented over such basis @xmath103 and the orbital populations are calculated as @xmath104 where @xmath105 is a matrix element of the overlap matrix . note that in principle the orbital population should be constructed for the atomic sic orbital @xmath72 . however , we notice that @xmath74 is rather insensitive to the specific choice of orbital , once this has a reasonable radial range . for practical numerical reasons in the present implementation , we always use the orbital populations projected onto the basis set sub - space consisting of the most extended first-@xmath99 orbitals of the atomic species involved . the matrix elements of the sic potential are simply @xmath106 and the asic - ks equation takes the final form @xmath107\psi_n^\sigma=\epsilon_{n}^{\sigma,\mathrm{sic}}\psi_n^\sigma\;. \label{ks - asic - final}\ ] ] with @xmath108 the pseudopotential . the energy corresponding to the sic - lsda functional is given by @xcite @xmath109=e^\mathrm{lsda}[\rho^\uparrow,\rho^\downarrow]- \sum_{n\sigma}^\mathrm{occ.}(u[\rho_n^\sigma]+e_\mathrm{xc}^\mathrm{lsda}[\rho_n^\sigma,0])\ ; , \label{totesic}\ ] ] where @xmath18=\int\mathrm{d}^3{\bf r}\:\frac{1}{2}\rho_n({\bf r})u([\rho_n ] ; { \bf r})\ ; , \label{hartree}\ ] ] @xmath110)=\int\mathrm{d}^3{\bf r}\:\rho_n({\bf r})\mathcal{e}^\mathrm{lsda}_\mathrm{xc } ( [ \rho_n ] ; { \bf r})\ ; , \label{xc}\ ] ] with @xmath111 the lsda exchange and correlation energy density . the orbital densities entering in the si term are those associated to the local orbitals @xmath37 . as we have already mentioned , this functional needs to be minimized with respect to the @xmath37 s , which are an implicit function of the total spin density @xmath6 . in the asic approximation these orbitals are not minimized , but taken as atomic functions . this means that in the present form the theory is not variational , in the sense that there is no functional related to the ks equation ( [ ks - asic - final ] ) by a variational principle . with this in mind we adopt the expression of equation ( [ totesic ] ) as a suitable energy , where the orbital densities are those given by the asic orbitals @xmath112 in our implementation the lsda ks energy @xmath113 is directly available as calculated in the _ siesta _ code @xcite and thus , only the second term of equation ( [ totesic ] ) needs to be calculated . this is easily done by calculating both @xmath0 and @xmath114 on an atomic radial grid for each atomic orbital in the system . we now compare our asic method with another atomic - like correction to lsda , namely the lda+@xmath0 method @xcite . in lda+@xmath0 one replaces the lsda exchange and correlation energy associated to the `` correlated '' orbitals ( @xmath13 or @xmath14 shells ) , with the hubbard-@xmath0 energy . thus the functional becomes @xmath115=e^\mathrm{lsda}[\rho({\bf r})]+e^\mathrm{u}[\{p_m^\sigma\}]-e^\mathrm{dc}[\{p_m^\sigma\}]\ ; , \label{ldau}\ ] ] where the hubbard energy @xmath116 and the double counting term @xmath117 depend on the orbital populations @xmath74 of the correlated orbitals . several forms for the lda+@xmath0 functional have been proposed to date . a particularly simple and transparent one @xcite , which is also rotationally invariant , redefines the @xmath0 parameter as an effective parameter @xmath118 and the functional takes the form @xmath119 \label{ldaurot}\ ] ] where in we separate out the index for the atomic position @xmath120 from the magnetic quantum number @xmath121 , and introduce the off - diagonal populations @xmath122 with @xmath123 . note that the lda+@xmath0 functional is si free for those orbitals that are corrected . although a rotationally invariant form of the asic potential can be easily derived , we assume here for simplicity that the system under consideration is rotationally invariant , or alternatively that we have carried out a rotation , which diagonalizes the @xmath124 matrix . in this case the energy becomes simply @xmath125\ ; , \label{ldaurotation}\ ] ] with @xmath126 . it is then easy to compute the ks potential @xmath127 \hat{p}^{i\phi}_m\ ; , \label{ldaupot}\ ] ] and the orbital energy @xmath128 these need to be compared with the asic potential ( equation ( [ psic - appr - final ] ) ) and orbital energy @xmath129 where the last term follows from @xmath130 and from equation ( [ totesic ] ) . the main difference between the asic and lda+@xmath0 method is in the way in which unoccupied states are handled . in fact , while lda+@xmath0 corrects unoccupied states and pushes the orbital energies upwards by @xmath131 , asic operates only on occupied states , that are shifted towards lower energies by @xmath132 . this reflects the fact that the sic is defined only for occupied ks orbitals . an important consequence is that the opening of bandgaps in the electronic structures , one of the main features of both the lda+@xmath0 and asic schemes , is then driven by two different mechanisms . on the one hand in lda+@xmath0 , gaps open up since occupied and unoccupied states are corrected in opposite directions leading to a gap of @xmath133 . on the other hand asic is active only over occupied states and gaps open only if occupied and unoccupied bands have large differences in their projected atomic orbital content . thus one should not expect any corrections for covalent materials where conduction and valence bands are simply bonding and antibonding states formed by the same atomic orbitals . this is for instance the case of si and ge . in contrast asic will be extremely effective for more ionic situations , where the orbital contents of conduction and valence bands are different . finally , by comparing the corrections to the orbital energy of a fully occupied state , one finds @xmath134 which establishes an empirical relation between the hubbard energy and the asic correction . since @xmath0 is sensitive to the chemical environment due to screening @xcite , while all the other quantities are uniquely defined by an atomic calculation , we can re - interpret the parameter @xmath71 as empirically describing the screening from the chemical environment within the asic scheme . the test calculations that we present in this work are for two classes of materials : extended and finite . first we investigate how our implementation performs in the solid state . in particular we discuss the rle of the parameter @xmath71 in determining the bandstructure of several semiconductors , considering both the ks band - gap and the position of bands associated with tightly bound electrons . the quasi - particle band gap @xmath135 in a semiconductor is defined as the difference between its ionization potential @xmath120 and electron affinity @xmath136 . these can be rigorously calculated from dft as the homo energy respectively of the neutral and negatively charged systems . this actual gap can not be directly compared with the ks band - gap @xmath137 , defined as the difference between the orbital energy of the homo and lumo states of the neutral system . in fact , the presence of a derivative discontinuity in the dft energy as a function of the electron occupation establishes the following rigorous relation @xcite @xmath138 with @xmath139}{\delta n}\left|_{n+\omega}-\frac{\delta e_\mathrm{xc}[n ] } { \delta n}\right|_{n-\omega}\right\}_{n_n}\ : . \label{derdisc}\ ] ] this is valid even for the exact xc potential , and therefore in principle one has to give up ks bandstructures as a tool for evaluating semiconductor band - gaps . the size of @xmath140 is however not known for real extended systems and the question of whether most of the error in determining @xmath135 from @xmath137 is due to the approximation in the xc potential or due to the intrinsic @xmath140 is a matter of debate . in general si - free potentials bind more than lsda and one expects larger gaps . surprisingly , functionals based on exact exchange , provide ks gaps rather close to the experimental values @xcite . the reason for such a good agreement is not fully understood , but it is believed that the exact ks gaps should be smaller than the actual ones . with this in mind , we adopt a heuristic approach and we use the ks band - gap as a quality indicator for interpreting the parameter @xmath71 and for providing its numerical value for different classes of solids . here we investigate the dependence of @xmath141 over @xmath71 and we determine the value for @xmath71 yielding the experimental band - gap . assuming that @xmath140 does not vary considerably across the materials investigated , this will allow us to relate @xmath71 to the degree of localization in a semiconductor and to extract the value useful for asic to be an accurate single - particle theory . in figure [ fig1 ] we present the band - gap of four representative semiconductors as a function of @xmath71 together with the value needed to reproduce the experimental band - gap . lsda corresponds to @xmath142 and while @xmath76 accounts for the full asic . calculated band - gap for nacl , mgo , aln and zno as a function of the parameter @xmath71 . @xmath142 is the lsda value and @xmath76 accounts for complete atomic sic . the lattice parameters used for the calculations are either the equilibrium lsda or the experimental when available.,scaledwidth=43.0% ] in general @xmath137 increases as @xmath71 increases , as a result of the stronger sic . the @xmath143 curve is almost linear with a slope , which appears to be material - specific . for the most ionic compound , nacl , the experimental gap is reproduced almost exactly by @xmath76 , i.e. by the full asic . this is somehow expected since the charge density of solid nacl is rather close to a superposition of the na@xmath144 and cl@xmath21 ionic charge densities . in this case of strongly localized charge densities the asic approximation is rather accurate yielding results substantially identical to those obtained with full self - consistent pz - sic @xcite . indeed earlier calculations for licl @xcite demonstrate that the sic band - structure is rather insensitive of the localized orbitals @xmath37 once these have an atomic - like form . for the other compounds the localized orbitals @xmath37 s are not necessarily atomic - like functions and deviations from @xmath76 are expected . interestingly we find that , for all the materials investigated , a value of around 1/2 reproduces the experimental band - gap rather accurately . as an illustration , in table [ tabgap ] we compare the experimental band - gap @xmath145 to the calculated @xmath137 for asic ( @xmath76 ) and lda , for several semiconductors ranging from ionic salts to wide - gap ii - vi and iii - v semiconductors . we also report the value of @xmath146 needed for @xmath147 . .[tabgap]experimental @xmath145 and ks @xmath137 band - gap ( in ev ) for a number of semiconductors . @xmath137 are calculated with both lsda and asic ( @xmath76 ) . in the last column we report the value of @xmath146 needed for @xmath147 . the lattice parameters used for the calculations are either the equilibrium lsda or the experimental when available ( in ) . rs = rocksalt , wz = wurtzite , zb = zincblende . the value for the experimental gaps are from the literature : @xmath148 @xcite , @xmath149 @xcite , @xmath150 @xcite , @xmath13 @xcite , @xmath151 @xcite , @xmath14 @xcite , @xmath152 @xcite , @xmath153 @xcite , @xmath154 @xcite [ cols="<,^,^,^,^,^",options="header " , ] again the asic@xmath155 results compare rather well with experiment , and we can conclude that the asic method offers a rather efficient and inexpensive theory for single particle vertical excitations . we are now in a position to discuss the homo - lumo gap in asic . as already mentioned , even for the exact xc functional , the ks gap @xmath156 does not account for the actual quasi - particle gap @xmath157 . this in turn is the sum of @xmath137 and the discontinuity of the exchange and correlation potential @xmath140 . equivalently @xmath158 i.e. @xmath140 is the discontinuity in the eigenvalue of the lumo state at @xmath3 . therefore , in order to extract the actual gap from the ks gap , provided that the spectrum is reasonably well described at integer electron numbers @xmath3 , what remains is to model the derivative discontinuity at @xmath3 and ensure that @xmath159 is relaxation free for ( @xmath160 ) . local and semi - local ( lsda / gga ) xc functionals lack such a discontinuity , while self - interaction corrections are able to restore it , at least in part . for instance the pz - sic scheme is successful in this regard . ionization curve for the ethylene ( c@xmath161h@xmath56 ) molecule as the occupation of the homo state is varied from 0 to 1 in going from the ionized c@xmath161h@xmath162 to the netural c@xmath161h@xmath56 state.,scaledwidth=45.0% ] in figure [ ethylene ] we illustrate the ionization curve for the ethylene ( c@xmath161h@xmath56 ) molecule as the occupation of the homo state is varied from 0 to 1 in going from the ionized c@xmath161h@xmath162 to the netural c@xmath161h@xmath56 configuration . it is seen that among the three schemes presented , only the pz - sic scheme approximately models the behaviour required by the equation ( [ discharge ] ) . the asic homo eigenvalue roughly agrees with the pz - sic eigenvalue at integer occupation but behaves linearly through non - integer values . thus we find that the derivative discontinuity for the molecule is smoothed out in asic , which still connects continuously different integer occupations . this is one of the limitations of the atomic representation employed in asic . in view of the foregoing discussion , the actual size of the homo - lumo gap in asic becomes significant with a direct bearing on the physics described . ideally , we want @xmath163 ( lumo for the @xmath3-electron system ) to be as close to @xmath164 so that the range of eigenvalue relaxation through fractional occupation numbers @xmath165 is minimized . looking at columns 6,7 and 8 in table [ eatab ] however , we see that for almost all the molecules , this is hardly the case . the agreement between @xmath163 and -ea from experiment ( @xmath166 ) for asic@xmath155 is quite poor implying a considerable energy range spanning fractional particle number . we still expect this energy range to be smaller for asic@xmath155 than lsda . it is also apparent from the table [ eatab ] that @xmath167 usually differs from @xmath168 and in fact by considerable magnitudes in some cases . thus the asic@xmath155 correction " to the empty lumo state does not vanish in contrast to the pz - sic scheme where , by definition , the empty eigenstates are sic free . since the sic operator @xmath169 is constructed in an atomic orbital representation , the correction to any ks eigenstate @xmath170 either filled or empty @xmath171 is not necessarily zero unless @xmath170 only projects onto empty atomic orbitals . also this correction to the lumo with respect to the lsda is negative in most cases , exceptions being nh@xmath161 and ch@xmath55 where it is desirably positive . thus the fundamental homo - lumo gap in asic is a combination of both the homo and lumo corrections . table [ holutab ] shows how this combination works out in asic@xmath172 and asic@xmath155 when compared to lsda . the molecular test set is the same as that in table [ iptab ] . lccc molecule & + + & lsda&asic@xmath172&asic@xmath155 + ch@xmath55 & 1.92 & 4.75 & 7.59 + nh@xmath173 & 7.1@xmath174 & 9.29@xmath174 & 11.61@xmath174 + sih@xmath175 & 8.44@xmath174 & 9.68 & 10.94 + c@xmath176h@xmath175 & 5.81 & 6.59 & 7.38 + sich@xmath175 & 6.19 & 7.07 & 8.06 + ch@xmath173chcl@xmath176 & 5.79 & 6.84 & 7.88 + c@xmath175h@xmath175s & 4.46 & 5.13 & 5.8 + c@xmath176h@xmath177s@xmath176 & 4.44 & 6.02 & 7.6 + pyridine & 3.85 & 4.56 & 5.26 + benzene & 5.22 & 5.9 & 6.59 + iso - butene & 4.88 & 5.56 & 6.26 + nitrobenzene & 3.25 & 4.03 & 4.42 + naphthalene & 3.36 & 3.83 & 4.29 + c@xmath178 & 1.62 & 1.87 & 2.12 + c@xmath179 & 1.75 & 1.99 & 2.23 + we see in almost all cases the asic gap is systematically larger than the lsda one . this is expected because the correction to the homo is usually much stronger than that to the lumo . in general , asic is expected to work well for systems where the occupied and un - occupied ks eigenstates of the extended system have markedly different atomic orbital signatures being derived predominantly from filled and empty atomic orbitals respectively . in such a case , the asic correction to the empty states would be nullified in being scaled by near - zero atomic orbital populations . in some cases , provided phase factors combine suitably , the correction to the empty states can even be positive with respect to the same in lsda . + before we conclude , we discuss some general properties of the asic method which are relevant to any orbital dependent sic implementation and also some possible pitfalls . as with other sic schemes , asic is not invariant under unitary transformations of the orbitals used in constructing the sic potential . thus the asic correction is likely to change as the atomic orbitals used for projecting onto the ks eigenstates of the system are rotated or transformed otherwise . unlike the perdew - zunger method however , there can be no variational principle over all possible unitary transformations of the atomic orbitals because in the general case they do not represent the hamiltonian of the system under consideration . this also implies that if the scheme is used with a system that is already well described by lsda , the correction " additional to the lsda result does not necessarily vanish . simple metals and narrow gap systems are likely candidates for this scenario . furthermore , its pertinent to mention that asic becomes ineffective if not counterproductive for materials with homonuclear bonding , in which valence and conduction bands have the same atomic orbital character . in this situation the asic potential will shift the bands in an almost identical way , without producing any quantitative changes , such as the opening up of the ks gap . note that this is a pitfall of the asic approximation , which distinguishes occupied from empty states only through their projected atomic orbital occupation , but not of the sic in general . typical cases are those of si and ge . the ks gap in si goes from 0.48 ev in lsda to only 0.57 ev for asic@xmath172 , while ge is a metal in both cases . in addition the lsda calculated valence bandwidths of 12.2 ev for si and 12.8 ev for ge , in good agreement with experiments , are erroneously broadened to 14.3 ev and 14.8 ev respectively . in conclusion , we have implemented the asic scheme proposed by filipetti and spaldin within the pseudopotential and localized orbital framework of the _ siesta _ code . we have then investigated a broad range of semiconductors and molecules , with the aim of providing a reasonable estimate for the scaling parameter @xmath71 . we found that @xmath76 , which accounts for the full atomic si , describes surprisingly well ionic semiconductors and molecules . in particular for molecules , both the ip and the ea can be obtained with good accuracy from the homo ks eigenvalues respectively for the neutral and singly charged molecule . this makes the asic scheme particularly suited for application such as quantum transport , where the position of the homo level determines most of the @xmath120-@xmath180 curve . in contrast iii - v and ii - vi semiconductors are better described by @xmath181 , which corrects the atomic si for screening . this makes asic@xmath172 an interesting effective band theory for semiconductors . the relation of the present scheme with the fully self - consistent sic methods has been emphasized , and so has been that with lda+@xmath0 . we acknowledge stimulating discussions with malgorzata wierzbowska , kieron burke and alessio filippetti . this work is supported by the science foundation of ireland under the grants sfi02/in1/i175 . dsp acknowledges financial support from the university of the basque country ( grant no . 9/upv 00206.215 - 13639/2001 ) , the spanish ministerio de educacion y ciencia ( grant no . fis2004 - 06490-c03 - 00 ) and the basque goverment and diputacion foral de giupuzkoa ( grant nanomateriales and nanotron , etortek program ) . traveling has been sponsored by enterprise ireland under the international collaboration programme ei - ic/2003/47 . computational resources have been provided by the hea iitac project managed by the trinity center for high performance computing and by ichec . temmerman , z. szotek and h. winter , phys . b * 47 * , 1184 ( 1993 ) ; z. szotek , w.m . temmerman and h. winter , phys . lett . * 72 * , 1244 ( 1994 ) ; a. svane , phys . rev . lett . * 72 * , 1248 ( 1994 ) ; a. svane , phys . b * 53 * , 4275 ( 1996 ) ; a. svane , v. kanchana , g. vaitheeswaran , g. santi , w.m . temmerman , z. szotek , p. strange and l. petit , phys . b * 71 * , 045119 ( 2005 ) . v. bougrov , m.e . levinshtein.,s.l . rumyantsev , a. zubrilov , in _ properties of advanced semiconductor materials gan , aln , inn , bn , sic , sige _ , eds . levinshtein m.e . , rumyantsev s.l . , shur m.s . , john wiley & sons , inc . , new york , 2001 , 1 - 30 .
we present an atomic orbital based approximate scheme for self - interaction correction ( sic ) to the local density approximation of density functional theory . the method , based on the idea of filippetti and spaldin [ @xcite phys . rev . b * 67 * , 125109 ( 2003 ) ] , is implemented in a code using localized numerical atomic orbital basis sets and is now suitable for both molecules and extended solids . after deriving the fundamental equations as a non - variational approximation of the self - consistent sic theory , we present results for a wide range of molecules and insulators . in particular , we investigate the effect of re - scaling the self - interaction correction and we establish a link with the existing atomic - like corrective scheme lda+@xmath0 . we find that when no re - scaling is applied , i.e. when we consider the entire atomic correction , the kohn - sham homo eigenvalue is a rather good approximation to the experimental ionization potential for molecules . similarly the homo eigenvalues of negatively charged molecules reproduce closely the molecular affinities . in contrast a re - scaling of about 50% is necessary to reproduce insulator bandgaps in solids , which otherwise are largely overestimated . the method therefore represents a kohn - sham based single - particle theory and offers good prospects for applications where the actual position of the kohn - sham eigenvalues is important , such as quantum transport .
several years ago , sergey fomin and andrei zelevinsky introduced a new mathematical object known as a cluster algebra which is related to a host of other combinatorial and geometric topics . some of these include canonical bases of semisimple algebraic groups , generalized associahedra , quiver representations , tilting theory , and teichmller theory . in the proceeding we will use the definitions and conventions used in fomin and zelevinsky s inital papers , @xcite . starting with a subset @xmath4 of cluster algebra @xmath5 , one applies binomial exchange relations to obtain additional generators of @xmath5 , called _ cluster variables_. the ( possibly infinite ) set of cluster variables obtained this way generate @xmath5 as an algebra . it was proven in @xcite and @xcite that any cluster variable is a laurent polynomial in @xmath4 , i.e. of the form @xmath6 where @xmath7 is a polynomial with integer coefficients ( not divisible by any monomial ) and the exponents @xmath8 are ( possibly negative ) integers . it is further conjectured that the polynomials @xmath7 have _ nonnegative _ integer coefficients for any cluster algebra . however , this conjecture has been proved in a limited number of cases , including the finite type case , as proved in @xcite , the case of rank two affine cluster algebras as demonstrated in @xcite , @xcite , @xcite , and @xcite as well as cluster algebras arising from acyclic quivers @xcite . the finite type case is defined as the case where the cluster variable generation procedure only yields a finite set of cluster variables associated to @xmath5 . by a combinatorial and geometric miracle , one which has sparked much interest in these algebras , the cluster algebras of finite type exactly correspond to the lie algebras of finite type . furthermore , in these cases , the cluster variables ( except for the @xmath9 s ) have denominators with _ nonnegative _ exponents and can be put in a @xmath10-to-@xmath10 correspondence with the positive roots of the associated root system . study of the particular finite type cluster algebra of type @xmath0 , also known as the ptolemy algebra has been especially fruitful as it can be realized in terms of the grassmannian and plcker embedding . in 2003 , as part of the reach research group under jim propp s direction , gabriel carroll and gregory price @xcite described two combinatorial interpretations of the associated cluster variables , one in terms of paths and one in terms of perfect matchings . further , ralf schiffler recently independently discovered and extended the paths interpretation @xcite . in the present paper , we go beyond @xmath0 , and describe a combinatorial interpretation for the cluster variables in all four families of finite type , namely @xmath0 , @xmath1 , @xmath2 , and @xmath3 , for the coefficient - free case . our combinatorial model will involve perfect matchings , in the spirit of @xcite , and agrees with carroll and price s interpretation in the @xmath0 case . unlike the aforementioned work we do not attempt to give the laurent expansion of cluster variables in terms of any seed but only in terms of the initial bipartite seed , whose definition we remind the reader of below . by restricting ourselves to expansions in this initial seed , we are able to explicitly write down families of graphs which encode the cluster algebra using weighted perfect matchings . we shall use the following notation throughout this paper . let @xmath11 be a finite graph with vertex set @xmath12 and edge set @xmath13 . for each edge @xmath14 , we set @xmath15 to be the weight of @xmath16 , where @xmath15 is allowed to be @xmath10 or @xmath9 for @xmath17 . a _ perfect matching _ @xmath18 of graph @xmath19 is a subset of @xmath20 such that for every vertex @xmath21 , there is exactly one edge @xmath22 containing @xmath23 . the weight of a perfect matching is defined to be the product @xmath24 , and we let @xmath25 denote the matching polynomial , or matching enumerator , of graph @xmath19 , defined as @xmath26 the result of this paper is the following theorem . [ vargraph ] let @xmath27 be a root system of classical type and denote its positive roots as @xmath28 . for each such @xmath27 , we explicitly construct a family of graphs , @xmath29 , with the following three properties . 1 . @xmath30 . 2 . for each @xmath31 , there exists a unique @xmath32 that can be effeciently identified . we have the cluster expansion formula @xmath33^{\phi } = \frac{p({g_\alpha}^{\phi})}{x_1^{\alpha_1}\cdots x_n^{\alpha_n}},\ ] ] where @xmath34^{\phi}$ ] denotes the cluster variable corresponding to positive root @xmath35 ( in type @xmath27 ) under fomin and zelevinsky s bijection . given graph @xmath36 , we are able to determine for which @xmath37 we have @xmath38 by breaking down @xmath19 into tiles . more precisely , we let a family of tiles @xmath39 be a finite set of graphs , with weighted edges , such that each @xmath40 is isomorphic to a cycle graph . given the faces and edge weighting of graph @xmath19 , we decompose @xmath19 into a union of such tiles by gluing together certain edges . we shall use the convention from @xcite , so that the initial exchange matrix @xmath44 contains rows of like sign . any rank @xmath45 cluster algebra of finite type has such a seed consisting of a cluster of initial variables @xmath46 and a set of @xmath45 binomial exchange relations of the form @xmath47 after mutating in the @xmath48th direction , i.e. applying an exchange relation of the form @xmath49 binomial , we obtain a new seed with cluster @xmath50 and exchange matrix @xmath51 such that the @xmath52 s satisfy @xmath53 as we mention below in remark [ sameseed ] , we shall use an ordering of mutations in this paper so that we need only work with binomial exchanges of the form @xmath54monomial @xmath55 . note that we shall use the notation @xmath56 to denote the numerator of the cluster variable with denominator @xmath57 despite its similarity with the notation of @xmath25 for the matching polynomial of graph @xmath19 . the outline of the paper is as follows . we proceed to prove theorem [ vargraph ] separately for the four families of non - exceptional type , starting with the well - studied case of @xmath0 . we will use different language than in @xcite , @xcite , or @xcite , and we include our own proof of this case to familiarize the reader with the techniques which we will utilize later in the paper . since the type of the cluster algebra will frequently be clear from context , we will simply denote tiles as @xmath40 or graphs as @xmath58 ( instead of @xmath59 ) . we end with some comments and directions for further research . in @xcite , fomin and zelevinsky explicitly constructed fibonacci polynomials for types @xmath0 and @xmath3 , which provide an alternate combinatorial expansion formula for cluster varibles . generalizations of these polynomials , for other types , are defined in @xcite , where they are referred to as f - polynomials . the work in this section was done independently of the work of carroll - price @xcite and the work of schiffler @xcite mentioned in the introduction . we will use the notation and the techniques of this section later in the paper for the @xmath1 and @xmath3 cases . thus we include this section even though the combinatorial interpretation given by proposition [ casean ] is not new in this case , although we believe our proof via excision , as described by lemma [ excisegr ] , is new . this excision technique will also be utilized for the @xmath1 and @xmath3 cases in section @xmath43 . we begin by reviewing the necessary characteristics of the cluster algebra of type @xmath0 . recall that lie algebra @xmath0 has a dynkin diagram consisting of a line of @xmath45 vertices connected by edges of weight one . @xmath60 thus the associated cartan matrix has the form @xmath61 and thus using the convention given in @xcite the associated exchange matrix is @xmath62 notice that every row has like sign and that the matrix is skew - symmetrizable ( and in fact skew - symmetric in this case ) . the bipartite seed for a cluster algebra of type @xmath0 therefore consists of a initial cluster of variables @xmath63 and exchange matrix @xmath64 which encodes the following exchange binomials @xmath65 we describe a set of tiles from which we will build our family of graphs . in the case of @xmath0 , let tiles @xmath66 be squares defined as follows : tile @xmath67 s northern edge is given weight @xmath68 while the other three are given weight @xmath10 . tile @xmath69 s southern edge is weighted with value @xmath70 and the rest are weighted with value @xmath10 . finally all other @xmath40 have a weight of @xmath71 given to their northern edge , @xmath72 for their southern edge while the eastern and western edges are given weight @xmath10 . + the tiles for cluster algebra of type @xmath73 . let @xmath74 be the set of graphs that can be built from these @xmath45 tiles given the following gluing rule . [ gluing ] without allowing reflections or rotations of the tiles , tile @xmath40 can be glued to tile @xmath75 if and only if the identified edge ( as an edge of @xmath40 ) lies clockwise from an edge weighted @xmath76 and clockwise from an edge weighted @xmath9 ( as an edge of @xmath75 ) . since tile @xmath40 only contains edges of weight @xmath71 and @xmath72 , and these weights appear across from each other , this rule uniquely describes how the blocks can connect . given the above tiles , @xmath77 , and the above gluing rule , the collection of possible graphs is enumerated by the set of subsets @xmath78 for @xmath79 . this collection @xmath74 has the same cardinality as the set of positive roots of the lie algebra of type @xmath0 using the bijection @xmath80 as shown in @xcite , this implies that the cardinality is also the same as the number of non - initial cluster variables for the bipartite cluster algebra of type @xmath0 . [ casean ] the set of graphs @xmath74 is in bijection with the set of non - initial cluster variables for a coefficient - free cluster algebra of type @xmath0 and satisfy the statement of theorem [ vargraph ] . + the collection @xmath81 . to prove this proposition we will take a detour through a case we refer to as @xmath82 . in this case , our set of tiles is in bijection with the integers , and we define @xmath40 to have @xmath71 on its northern edge , @xmath72 on its southern edge for all @xmath83 . without the issue of boundaries , it is easier to show that a certain lattice of graphs corresponds to the non - initial cluster variables . after doing so , we choose a periodic specialization for the initial variables to recover a corresponding region of this lattice for any specific @xmath0 . we start with the following observation . [ sameseed]for all @xmath45 , if we start with the above exchange matrix @xmath64 and apply the binomial exchanges corresponding to relations @xmath10 , @xmath84 , @xmath85 ( resp . @xmath86 ) if @xmath45 is odd ( resp . even ) the resulting exchange matrix is @xmath87 . afterward , applying the relations @xmath88 , @xmath43 , @xmath89 ( resp . @xmath45 ) if @xmath45 is odd ( resp . even ) to exchange matrix @xmath90 results in the intial exchange matrix @xmath64 . in fact , in both of these cases , the order of the exchanges does not matter , and the intermediate exchange matrices will have rows of like sign for all relevant @xmath91 not already exchanged . by the definition of matrix mutation , this procedure will in fact work for any cluster algebra where the seed has an exchange matrix that is tri - diagonal ( @xmath92 if @xmath93 ) . thus we can calculate a row of cluster variables at a time by applying the exchange relations relative to the two previous rows . the tri - diagonal condition includes the cases @xmath0 , @xmath1 , @xmath2 , and @xmath94 and minor modifications to the procedure will allow it to work for @xmath3 . returning to the @xmath0 case , after applying exchange @xmath95 ( resp . @xmath86 ) we have cluster @xmath96 @xmath97 where @xmath98 using the convention @xmath99 . analogously , applying exchanges @xmath100 ( resp . @xmath45 ) we obtain the cluster @xmath101 @xmath102 where @xmath103 for @xmath104 , if we set @xmath105 . by remark [ sameseed ] we can make a lattice of cluster variables by applying exchanges iteratively ( row - by - row ) in this order . @xmath106 + top six rows of this lattice ( assuming @xmath45 odd ) . by the binomial exchange relations , this lattice satisfies the diamond condition which states that the relation @xmath107 holds for any four elements arranged as a diamond . @xmath108 for example @xmath109 for @xmath110 . furthermore , this lattice can be extended periodically by using the conventions @xmath105 , @xmath111 , and extending further using negatively weighted variables . note that the negatives are necessary since we wish the configurations @xmath112 to satisfy the diamond condtion . @xmath113 + six rows of extended lattice ( assuming @xmath45 odd ) . this lattice can also continue infinitely in the vertical direction , as well as horizontally , extending vertically in the unique way that preserves the diamond condition throughout the entire lattice . consequently all @xmath0 can be treated simultaneously by considering the infinite diamond pattern ( @xmath82 ) which starts with sequence @xmath114 zig - zagging to create the initial two rows . to obtain the extended @xmath0-lattice for a specific @xmath45 we let @xmath115 and then take the limit as @xmath116 goes to zero . we do not set @xmath116 and @xmath117 to be zero directly since this would sometimes result in indeterminate expressions of the form `` @xmath118 '' . also we use the shifted indices here since it will make the ensuing arguments more symmetrical . as a consequence of these substitutions , it suffices to start the proof of proposition [ casean ] by proving the combinatorial interpretation for the infinite diamond pattern corresponding to @xmath82 , which we write in terms of @xmath119 s for @xmath120 . even though we now have a boundary - less lattice , every given @xmath121 can be computed locally by considering the necessary mutations stemming from a finite half diamond extending back to the initial two rows of @xmath119 s . + the tiles for the extended @xmath0-lattice . for the purposes of the @xmath122 case , for all @xmath123 , we let @xmath124 denote the tile with @xmath125 on its northern edge and @xmath126 on its southern edge . we will utilize variables @xmath119 s and tiles @xmath124 s until we do the final substitution at the end of the proof of proposition [ casean ] . we now proceed to prove the combinatorial interpretation in this boundary - less version . we start with the base case where we can easily see that the combinatorial interpretation works for cluster variables with denominator @xmath119 . to see this , we observe that @xmath127 corresponds to the two perfect matchings of the graph consisting of tile @xmath128 by itself . similarly , we observe the bijection works for the second row of non - initial cluster variables by the definition of @xmath129 . we see this by verifying that the graph containing tiles @xmath130 , @xmath131 , @xmath132 connected in that order bijects to cluster variable @xmath129 , i.e. @xmath133 . by a technique of graphical condensation developed by eric kuo @xcite , we obtain the following combinatorial interpretation for the rest of the rows . [ diag ] cluster variable @xmath121 bijects to graph @xmath134 i.e. the grid graph containing exactly @xmath135 tiles , namely tiles @xmath136 through @xmath137 connected in order . the proof follows from a slight variant of the argument given in @xcite . here we need to be more careful with the labeling scheme , but the same pairings will yield the desired result . in fact if one lets @xmath138 if @xmath139 even and @xmath140 if @xmath139 odd , one recovers the @xmath141 case analyzed in @xcite . in particular , we inductively assume for all @xmath123 that @xmath142 bijects to graph @xmath143 and @xmath144 bijects to graph @xmath145 , in the sense that @xmath146 and @xmath147 . it thus suffices to show , for all @xmath123 , that laurent polynomials @xmath121 , defined as @xmath148 , equal @xmath149 where @xmath150 we use our induction hypothesis and normalize to rewrite our desired equation as @xmath151 one can decompose graph @xmath152 into a superposition of graphs @xmath153 so that @xmath154 is the intersection of overlap . out of the two subgraphs , only @xmath155 contains tiles @xmath156 and only @xmath157 contains @xmath158 . let @xmath159 denote the set of perfect matchings of graph @xmath19 , @xmath160 denote the matching of @xmath152 using the horizontal edges of @xmath161 , and @xmath162 denote the matching of @xmath154 using the horizontal edges of @xmath163 . the pair of matchings @xmath164 has exactly the weight of the excess monomial @xmath165 we finish the proof of lemma [ diag ] by exhibiting a weight - preserving bijection between @xmath166 and @xmath167 , thus showing ( [ norma ] ) . we define our bijection piece - meal on @xmath168 , first considering the case where the horizontal edges of penultimate tile @xmath169 in @xmath152 are not used . in this case , the pair of matchings from @xmath170 reduces to a pair from @xmath171 . we define @xmath172 for such matchings by letting @xmath173 be the corresponding matching of @xmath155 and build matching @xmath174 by adjoining the matching of @xmath154 to the matching of @xmath175 . in other words , map @xmath176 takes tile @xmath175 and slides it down from @xmath152 onto @xmath154 to obtain @xmath157 with the matching included . this also leaves @xmath155 in place of @xmath152 . if on the other hand , the horizontals of @xmath169 _ are used _ , then the situation is more complicated . if we restrict further to the case where the rightmost vertical edge is used in @xmath154 , we can define @xmath176 analogously by sliding down tiles @xmath169 and @xmath175 . we are also forced to use the rightmost vertical edge of @xmath155 in this case . we can continue defining @xmath176 iteratively , defining it for classes characterized by the length of the pattern of horizontals on the right - hand sides of @xmath152 and @xmath154 . if the horizontals of @xmath177 in @xmath152 , the horizontals of @xmath178 @xmath179 ( resp . @xmath180 ) in @xmath154 are used , accompanied by a vertical edge between tiles @xmath181 and @xmath182 of @xmath152 ( resp . @xmath182 and @xmath183 of @xmath154 ) , then @xmath176 swaps the right - hand sides of these two graphs , leaving the left - hand sides alone up until tile @xmath184 ( resp . @xmath181 ) . this construction makes sense as long as we eventually encounter a vertical edge as we move leftward via the use of these horizontal edges since given such patterns , neither the matching of @xmath152 nor of @xmath154 , will use the horizontal edges of tile @xmath181 ( resp . @xmath182 ) . map @xmath176 is injective since the inverse map just swaps back the right - hand sides as dictated by the alternating pattern of horizontals . since we have exhaustively enumerated the pairs of matchings @xmath185 by splitting into classes according to the longest alternating pattern of horizontals , we also have surjectivity . lastly , it is easy to verify that @xmath186 is the unique matching which can not be decomposed into a pair @xmath185 . analogous pairings will also appear below in the arguments for the case of @xmath1 . notice that by lemma [ diag ] , the diagonals of the lattice satisfy the following two properties : @xmath187 on any of the diagonals travelling from _ sw to ne _ , all graphs _ end _ with the same tile . @xmath187 on any of the diagonals travelling from _ nw to se _ , all graphs _ start _ with the same tile . we now wish to show how to specialize to the case of a specific @xmath0 by imposing periodicty and boundary conditions on the intial two rows of variables . to accomplish this goal , we must ( a ) verify that the @xmath121 s satisfy the correct horizontal periodicity of the extended-@xmath0 lattice once we apply the proper substitutions of variables , ( b ) verify that we have a vertical periodicity as well and really only need to worry about a finite collection of graphs which we readily identify as the set @xmath74 . to show ( a ) , it suffices to show @xmath188 , @xmath189 , and @xmath190 for all @xmath191 . the diamond condition will then induce the horizontal periodicity for non - initial cluster variables . notice that @xmath192 and the periodicity @xmath193 imply the relation @xmath194 for all @xmath195 . thus lemma [ diag ] , together with @xmath196 , imply the relations @xmath197 where the second equality comes from reversing the order of the tiles . furthermore , tile @xmath198 has @xmath199 on its northern edge and @xmath200 on its southern edge , while @xmath201 has @xmath202 on its northern edge and @xmath203 on its southern edge . thus the relation @xmath196 for all @xmath123 induces @xmath204 for all @xmath120 and @xmath205 . as a corollary , we obtain @xmath206 for all @xmath191 . under lemma [ diag ] , the graph corresponding to @xmath207 is centered around tile @xmath208 , i.e. @xmath209 laurent polynomial @xmath210 analogously corresponds to a graph centered around tile @xmath211 , which is equivalent to tile @xmath208 after applying the above periodicity and substution in the extended @xmath0-lattice . consequently , we induce that for all @xmath205 , @xmath212 similarly . to complete our proof of ( a ) , we prove the following lemma . [ centerone ] if @xmath213 signifies the graph @xmath214 , i.e. a graph with an even number of tiles , with leftmost tile @xmath215 , and centered around subgraph @xmath216 , then ( using @xmath196 ) , after dividing the matching polynomial @xmath217 by the proper monomial , we find that graph @xmath213 bijects to laurent polynomial @xmath218 . moreover , any graph @xmath219 with an odd number of tiles centered around tile @xmath220 ( resp . @xmath221 ) bijects to @xmath10 as a laurent polynomial , regardless of the number of tiles . we start by proving the result for graphs with an even number of tiles , beginning with two base cases . the graph @xmath216 has three perfect matchings , but due to signs , two of them cancel with one another . we thus obtain @xmath222 . after dividing through by @xmath223 we get laurent polynomial @xmath224 . secondly , the graph @xmath225 has eight perfect matchings , but one of them has a weight containing submonomial @xmath226 , and six of the other seven cancel with each other after letting @xmath196 and @xmath227 . we are left with one perfect matching , which has weight @xmath228 . this time the denominator is @xmath229 and we get @xmath230 after division . we now assume that @xmath231 and wish to show @xmath232 for all such @xmath191 . notice that if the rightmost vertical edge of graph @xmath233 is used in a matching , our computation of @xmath217 reduces to the computation of @xmath234 , where @xmath235 , a graph centered around tile @xmath208 . however , we know from previous arguments that such a graph corresponds to zero as a laurent polynomial . thus any matching with the rightmost vertical edge of graph @xmath236 does not contribute to the enumerator @xmath217 . thus we must use the two rightmost horizontal edges , which have weight @xmath237 , and then compute @xmath238 where @xmath239 . however , such a graph is the horizontal reflection of graph @xmath240 and so by analogous logic , we are now forced to use the two _ leftmost _ horizontal edges in our matchings to get a nontrivial contribution . such edges have weight @xmath241 . consequently , after two iterations , and two helpfully placed negative signs , we have the identity @xmath242 induction thus yields the result for the case of @xmath213 , i.e. a graph with an even number of tiles centered around @xmath216 . we now prove this result for the corresponding graphs with an odd number of tiles . we only describe the proof for the case of those centered around @xmath220 since the proof for those centered around @xmath221 is analogous , only with messier notation . we let @xmath243 and for the moment ignore boundaries and periodicity ; only use the assignments @xmath227 , @xmath244 , and take the limit of @xmath245 as @xmath246 . we look at possible perfect matchings of graph @xmath219 , and note that if both horizontal edges or both vertical edges of tile @xmath247 are used , then we are reduced to computing @xmath248 with @xmath249 . however , since @xmath250 is centered around tile @xmath208 , the laurent polynomial @xmath251 tends to zero as @xmath246 . we conclude that any matching of @xmath219 resulting in a nontrivial contribution to @xmath252 must utilize the two horizontal edges of tile @xmath253 with weight @xmath237 . however , this step reduces our calculation to that of the matching polynomial of @xmath254 , which has an even number of tiles and is centered around @xmath255 . by our earlier logic , we therefore can only have a nontrivial contribution to @xmath256 in the case where we use the two leftmost horizontal edges , which have weight @xmath257 . this reduction results in subgraph @xmath258 , which is centered around @xmath259 , and so by induction @xmath260 which is the same as the denominator corresponding to @xmath243 . we thus have shown @xmath261 , i.e. that after applying substitutions to the initial row of the @xmath122 lattice , we get the extended @xmath0 lattice with the proper horizontal periodicity . we now wish to show the vertical periodicity of @xmath262 . [ excisegr ] whenever a graph contains the tile @xmath220 ( resp . @xmath221 ) , we may excise the graph by removing a subgraph centered around @xmath220 ( resp . @xmath221 ) . without changing the corresponding laurent polynomial . underneath initial variables @xmath263 , any graph containing tile @xmath208 ( resp . @xmath211 ) , or other forbidden tiles @xmath264 with @xmath139 outside the range @xmath265 must contain either tile @xmath220 or tile @xmath221 . consequently , this process of excision is sufficient to eliminate all tiles @xmath264 with @xmath266 . ( we will see shortly that the apparent problem of a graph containing both @xmath220 and @xmath221 where the subgraphs of excision overlap is not actually an issue . ) without loss of generality , assume that graph @xmath19 contains tile @xmath220 and has the form @xmath267 for @xmath268 . the content of the claim is that graph @xmath19 and graph @xmath269 biject to the same cluster variable . we categorize perfect matchings of @xmath19 based on whether or not two horizontal edges appear on the tile @xmath253 . if they do not , then that matching of @xmath19 decomposes into a matching of subgraph @xmath270 , and a matching of @xmath271 . thus we obtain @xmath272 as a contribution to the laurent polynomial corresponding to graph @xmath19 , where the second equality follows from lemma [ centerone ] . if on the other hand , the horizontal edges of @xmath253 _ are used _ , the matching decomposes differently and we get a contribution of @xmath273 which equals @xmath274 by lemma [ centerone ] . comparing the sum of these two laurent polynomials to the cluster variable corresponding to graph @xmath275 finishes the proof . with lemma [ excisegr ] proved , we turn our attention back to the proof of proposition [ casean ] . [ proof of proposition [ casean ] ] recall that the extended @xmath0-lattice is horizontally periodic . we thus restrict our attention to a region that lies between the two columns of positive ones and below the strip of initial variables . @xmath276 + the first twelve rows of this region for @xmath277 . cluster variable @xmath121 contains tiles @xmath278 through @xmath175 for a total of @xmath135 tiles . since we are considering only cluster variables between the two columns of positive ones , we have graphs centered at @xmath264 for @xmath279 . thus our graph either \1 ) contains no forbidden tiles . + 2 ) contains forbidden tiles including @xmath220 and tiles to the left . + 3 ) contains forbidden tiles including @xmath221 and tiles to the right . + 4 ) contains both sets of forbidden segments . if there are forbidden tiles on only one side , then lemma [ excisegr ] allows us to edit out the forbidden segment . however , if there are forbidden strips on both sides we would encounter a problem if the segments we were deleting overlapped . so consider a graph that contains tiles @xmath280 , along with @xmath281 tiles to the left and @xmath282 tiles to the right . by lemma [ excisegr ] , when we get rid of the @xmath283 tiles on the two ends , we would also be deleting @xmath284 tiles from the middle ( including @xmath220 and @xmath221 ) . for the first @xmath285 rows of our lattice , the number of total tiles is @xmath286 and thus @xmath287 which implies that @xmath288 . thus there can not be an overlap in these first @xmath285 rows . furthermore , recalling our indexing @xmath289 , we get as an added bonus that the graph corresponding to @xmath290 , which consists of @xmath291 tiles , bijects to the same cluster variable as the graph of the single tile @xmath292 ( resp . @xmath293 ) for @xmath139 even and @xmath45 even ( resp . @xmath45 consequently , we get a combinatorial interpretation for the first @xmath285 rows . we will refer to the region of the extended @xmath0-lattice that lies between the two columns of positive ones and in the first @xmath285 rows underneath the initial variables as the @xmath0-lattice . the diagonals of the @xmath0-lattice inherit the properties of the @xmath122-lattice and this implies that all the cluster variables do in fact appear in our lattice . in particular , their denominators are in bijection with positive roots of @xmath0 s root system and we have the desired corresponding graph for each of them . we now can complete the substitutions by letting @xmath294 , for @xmath295 , which induces @xmath296 for @xmath297 . in particualr , the extended @xmath0 lattice reduces to doubly - periodic copies of the @xmath0 lattice containing graphs involving only tiles @xmath298 . thus propositon [ casean ] is proven . as a corollary of this argument , the diamond condition implies that the @xmath299nd and @xmath300rd rows consists of the initial cluster variables written in reverse order . consider a new lattice @xmath301 consisting of connected subsets of @xmath77 such that @xmath302 appears in the graph associated to @xmath303 and add columns consisting of empty sets on the left - hand and right - hand sides of this lattice . this lattice satisfies a tropical - like diamond condition where one of the four following hold . such lattices are known as frieze patterns , and were studied by conway and coxeter @xcite in the 1970 s . such patterns have also been studied in connection with cluster algebras in work of caldero @xcite and work of propp @xcite . these lattices are also special cases of the bipartite belt described in @xcite ; each row of the lattice corresponds to a seed of the belt . hugh thomas @xcite brought it to the author s attention that one can also derive the above lattices via the algorithm for constructing the auslander - reiten quiver @xcite starting from projective representations ; in particular the pattern of denominator vectors agrees with the dimension vectors of the indecomposables in the ar quiver . the lie algebra @xmath2 has the following dynkin diagram @xmath305 and thus the bipartite exchange matrix is : @xmath306 to build the corresponding graphs we let @xmath307 be identical to @xmath77 except that tile @xmath67 now has weights of @xmath68 and @xmath68 opposite each other instead of a lone weighted edge . this change to @xmath67 corresponds to the change to the exchange polynomial associated to label @xmath10 in the seed of this cluster algebra . + tiles for @xmath308 . we use gluing rule [ gluing ] again which leads us to a collection similar to @xmath74 except now tile @xmath67 can connect to tile @xmath309 on either side . thus the collection of possible graphs , @xmath310 corresponds to the sets of the form @xmath311 for @xmath312 or multisets of the form @xmath313 for @xmath314 . this collection @xmath310 has the same cardinality as the collection of non - initial cluster variables for a cluster algebra of type @xmath2 and thus the collection of positive roots for a root system of type @xmath2 , as in the last case @xcite . [ casecn ] the set of graphs @xmath310 is in bijection with the set of non - initial cluster variables for a coefficient - free cluster algebra of type @xmath2 such that the statement of theorem [ vargraph ] holds . this can be proved quickly by using the folding procedure as in @xcite . we identify @xmath315 with @xmath2 by letting @xmath316 for @xmath317 . we let @xmath318 and let @xmath319 for @xmath320 . our lattice will contain repeats but we can restrict our list to the right half , including the central axis , to obtain the correct number of graphs . thus proposition [ casean ] implies proposition [ casecn ] . + the collection @xmath321 with duplicates . in the previous two cases , all of the exchange polynomials had degree two or less . for the cases of @xmath1 and @xmath3 , exactly one of the exchange polynomials has degree three . we will deal with such exchanges by including hexagons as potential tiles . we start with the case of @xmath1 , which is a folded version of the simply - laced @xmath3 case . by folding , our proofs will require less notation and as we will see , the @xmath3 case has a symmetry such that we can easily derive this case from the results for @xmath1 . in the case of @xmath1 , the dynkin diagram is @xmath322 and thus the bipartite exchange matrix is : @xmath323 we will now use the notation @xmath67 through @xmath69 to refer to a collection of tiles , @xmath324 , related to @xmath1 . we construct @xmath324 from @xmath77 by first replacing @xmath309 with a hexagon having weights @xmath10 , @xmath325 , @xmath10 , @xmath325 , @xmath10 , and @xmath326 in clockwise order starting from the top . we let @xmath67 be a trapezoid with a single weighted edge of @xmath68 on its northern side . note that @xmath67 is homeomorphic to its previous definition . then for all @xmath327 we define @xmath40 , for type @xmath328 , as a counter - clockwise rotation of the @xmath0-tile @xmath40 , including the boundary tile @xmath69 which have a single weighted edge of @xmath70 on its eastern side . + tiles for @xmath329 . the gluing rule will be more complicated now that hexagons are involved . as a first approximation , the set of graphs @xmath330 will include any graphs that can be constructed from @xmath324 while conforming to rule [ gluing ] . again we are not allowing rotation or reflections of the tiles so they must be connected in the orientaitions as described above . any such graph will resemble either a _ tower _ of tiles @xmath331 through @xmath332 for @xmath333 , a _ base _ involving hexagon @xmath309 with or without trapezoid @xmath67 on its western side , or may be a complex of a tower beginning with @xmath334 on top of a base . in addition , we enlarge the set of @xmath330 by allowing any graphs that obey the following second rule : [ hexagons ] the trapezoidal tile @xmath67 may appear twice if and only if the lift of the graph to @xmath335 ( i.e. @xmath45 arbitrarily large ) has one of the following three forms : where @xmath336 and @xmath337 are both odd . notice that rule @xmath10 is now broken when we connect trapezoid @xmath67 to a hexagon @xmath309 on its left . furthermore , in the last of these cases , we have adjoined an additional arc which had not been allowed or required in previous examples . however , there is precedent for using such additional arcs , see section @xmath84 of @xcite . it will develop that we can project these graphs , consisting of two towers , down to the @xmath1-lattice by excision around tile @xmath338 just as in the @xmath0 case . thus the fact that these are an odd number of tiles in each tower , with the larger tower on the left , greatly limits the set of such graphs . one can check that the collection of graphs @xmath330 obeying rule @xmath10 _ or _ rule @xmath88 has the cardinality equal to the number of positive roots for @xmath1 . further we will prove , after excision , that theorem [ vargraph ] is satisfied by these defintions . [ casebn ] the set of graphs @xmath330 is in bijection with the set of non - initial cluster variables for a coefficient - free cluster algebra of type @xmath1 such that the statement of theorem [ vargraph ] holds . analogous to the @xmath0 case , we will first prove the result for the @xmath339 case , that is we assume that @xmath45 is arbitrarily large so that for @xmath340 tile @xmath40 always has exactly two weighted edges ( @xmath72 on its east and @xmath71 on its west ) . this greatly simplies the proofs by allowing easier notation and bypassing case - by - case analysis . we will later discuss how to obtain the result for a specific @xmath1 from such graphs . we create a semi - infinite lattice whose entries satisfy a deformed diamond condition . @xmath341 + the lattice for @xmath339 . without the boundary on the right , any collection of four variables @xmath108 such that @xmath342 and @xmath343 will satisfy @xmath344 . a diamond such that @xmath345 will satisfy the truncated condition @xmath346 , and a diamond which contains @xmath347 will satisfy the relation @xmath348 . as before , we say that a laurent polynomial bijects to graph @xmath19 , which we denote as @xmath349 , if @xmath350 where @xmath25 is the matching polynomial of @xmath19 and @xmath351 encodes the number of occurences of tile @xmath40 in graph @xmath19 . given this setup along with the initial assignments of @xmath352 for @xmath353 , we directly verify that @xmath354 , @xmath355 , + @xmath356 , @xmath357 , where the weights of the edges are as dictated by the definitions of tiles @xmath67 through @xmath69 . additionally for @xmath358 , the only initial variables used to determine @xmath303 are @xmath359 and thus we recover the regular diamond pattern used in the @xmath122 case . consequently , we immediately obtain [ aaregion ] @xmath360 for @xmath361 when @xmath362 . we also find that @xmath363 . we proceed with the rest of the proof in three steps . the first two steps are proved inductively by using lemma [ aaregion ] as well as @xmath364 as a base case . we will prove the inductive step via the usual diamond condition @xmath344 which will hold for the diagonals @xmath365 and @xmath366 while @xmath367 . [ xii ] @xmath368 for @xmath367 . [ xii+2 ] @xmath369 for @xmath367 . the proof of these lemmas will prove proposition [ casebn ] for all @xmath303 such that @xmath370 . we must now use variants of the diamond conditions ( @xmath346 and @xmath371 ) to extend down columns @xmath372 and @xmath373 respectively . but to continue to have new entries to use as @xmath374 in the relation we must continually extend down diagonals as we extend down the columns . consequently , we proceed to prove the following three results for @xmath375 , then for @xmath376 , and so on by induction . [ restofdiagonals ] @xmath377 for @xmath378 , + @xmath379 for @xmath380 , + @xmath381 where @xmath382 for @xmath367 and @xmath378 . + a model of how these lemmas fit together and relate to the @xmath339-lattice . with the proof of proposition [ casebn ] now broken down into manageable chunks , we proceed to prove the lemmas . * proof of lemma [ xii ] . by lemma [ aaregion ] , we have that the northeast portion of the lattice is filled in , and we use the entries on diagonal @xmath383 and the base cases of @xmath384 and @xmath385 to extend to the rest of diagonal @xmath386 by the diamond condtion . assuming that cluster variables are * @xmath387 + @xmath388 + @xmath389 we wish to show that @xmath390 , given the diamond relation @xmath391 . first of all , we see that the occurrences of tiles match up on each side of the equal sign , which implies that the denominators agree appropriately . it suffices to show the weighted number of matchings also match up accordingly . any pair of matchings of the graphs and can be decomposed into a pair of matchings on graphs and except for one matching . the logic is identical to that of lemma [ diag ] . here we swap the tops of the towers and note if the top horizontal edge of the hexagon is used ( instead of the nw and ne diagonal edges ) , then completely swapping the two towers is permissible . this extraneous indecomposable pairing is the pair and it has exactly the correct weight @xmath392 and thus the lemma is proved . * proof of lemma [ xii+2 ] . the proof of this lemma is analogous except that we shift the diamond pattern so that * @xmath387 , @xmath390 + @xmath388 , @xmath389 . again , we have a bijection by swapping the tops of the ( left ) towers , and there is exactly one extraneous pair of matchings : . this pair has precisely the correct weight of @xmath393 . * proof of lemma [ restofdiagonals ] . the first part is proven by the following observations . if we * let graph @xmath394 be , @xmath94 be , @xmath67 be , @xmath309 be , @xmath395 be , and let @xmath396 be then @xmath397 putting these two equalities together we obtain @xmath398 most matchings of @xmath396 correspond to a matching of @xmath395 by the usual procedure of swapping the right - hand sides . the extraneous matching of @xmath396 has the form , contributing a factor of @xmath399 , and yielding the identity @xmath400 since @xmath401 is satisfied by letting @xmath402 and @xmath403 , part one of lemma [ restofdiagonals ] is proved . part three is proved analogously to lemma [ xii+2 ] . in this case , we have a diamond where all four entries are graphs consisting of two towers on the maximal base of two trapezoids and two hexagons . we inductively know the validity of these graphs for laurent polynomials @xmath281 , @xmath282 , and @xmath374 , so it is sufficient to verify the diamond condition if @xmath404 . for ease of notation , we temporarily let @xmath405 , @xmath406 , @xmath407 , and @xmath408 be the graphs corresponding to these particular laurent polynomials . as before , we wish to present a bijection @xmath409 where @xmath410 is a specific pair of matchings . map @xmath176 starts by swapping the two left towers of @xmath405 and @xmath408 if able ; this is analogous to the earlier cases . however , in the case where these towers can not be swapped ( because the alternating pattern continues down into the base ) map @xmath176 then attempts to swap the right towers of @xmath405 and @xmath408 . there is exactly one pair of matchings where both attempts at swapping fails . this is the pair of matchings where the alternating patterns appear on both towers down through the bases ; by inspection , such a pair has precisely the weight of the extraneous monomial . this leaves part two as the crux of the proof and the step which utilizes the diamond relation @xmath348 which makes @xmath339 different from the previous cases . we wish to show that the assignments @xmath387 , @xmath390 + @xmath388 , @xmath389 satisfies @xmath411 . we shall use the fact that if @xmath388 , then @xmath412 . clearly the denominator corresponds correctly . the number of weighted matchings , and thus the numerator , is also correct since there is a weight - preserving bijection between matchings of @xmath413{lem8b2pf.eps } \hspace{0.1in}$ ] and matchings of @xmath413{lem8b2pf2.eps}$ ] . with this substitution , a superposition argument analogous to that which just proved part three demonstrates the validity of the standard diamond relation @xmath414 . with the last step completed , these three lemmas prove that the cluster variables of the @xmath339-lattice correspond exactly to the desired graphs . there is a pattern inherent in the nw to se and ne to sw diagonals once again . this time , this pattern manifests itself ( in the region where @xmath415 ) by dictating the choice of right tower ( nw to se ) and the choice of left tower ( ne to sw ) . recall that lemma [ aaregion ] already described the pattern in the region where @xmath416 using grid graphs consisting of tiles @xmath417 where @xmath418 . we now turn to the problem of restricting to specific @xmath1 . we use the same methodology as in the @xmath0 case . given that we did not include the right - hand boundary , graphs can contain tile @xmath40 for arbitrarily large @xmath139 . we thus want to essentially let @xmath419 to force tile @xmath69 to have the proper weights , as a tile in @xmath324 as opposed to @xmath420 . however , unlike the @xmath0 case , we can not simply apply this substitution and use horizontal periodicity to make sure the diamond condition holds throughout . the problem is that the left - hand boundary satisfies a diamond conditon of a different form . nonetheless , the diagonals to the northeast , those where @xmath362 , contain graphs which are connected grid graphs , towers , of the form @xmath421 where @xmath363 , and any neighboring four entries satisfy the same diamond condition as the @xmath0 case . thus the logic of lemma [ centerone ] carries over and we can excise connected subgraphs centered at tile @xmath338 ( @xmath422 under the old notation ) . in particular , we obtain the following excision . if @xmath282 satisfies @xmath423 then the graphs @xmath424 and @xmath425 biject to the same laurent polynomial . this lemma follows directly from lemma [ centerone ] after replacing @xmath338 by the equivalent tile @xmath422 . the restriction of @xmath426 must be added here since we have a boundary on the left side , i.e. hexagon @xmath309 can not be excised during this procedure . notice that at the extreme , @xmath427 bijects to the same laurent polynomial as @xmath334 , and further @xmath428 is centered around @xmath338 and thus bijects to the laurent polynomial @xmath10 , the same as the empty graph . using this lemma , we are able to determine the northeast corner of the finite @xmath1 lattice . using these graphs during the inductive step of lemma [ xii ] in lieau of the arbitrarily large towers of @xmath429 allows us to fill in the next diagonal of the @xmath1 lattice where the towers sitting on the base of @xmath430 will consist exclusively of the tiles between @xmath334 and @xmath69 . simiarly , the recursive steps of lemmas [ xii+2 ] and [ restofdiagonals ] also follow with these truncated graphs , with tiles between @xmath67 and @xmath69 used instead . applications of lemma [ xii+2 ] and then successive applications of the third part of lemma [ restofdiagonals ] will allow this interpretation to extend down sw to all but a se corner of the lattice . note that the diagonals again determine the left - hand and right - hand towers of the @xmath1-lattice since this property is inherited from the @xmath339-lattice . we compute the se corner by starting at the bottom with initial row @xmath309 , @xmath431 , @xmath432 , @xmath433 , and propogating _ upwards _ via the diamond condition . we get via lemma [ aaregion ] , now with tile @xmath309 , instead of @xmath334 , as the smallest allowable tile , all but a single diagonal of the se corner . the final diagonal has the form @xmath434 , proven by applying lemma [ xii+2 ] upwards . notice that in the end , we obtain a lattice where the nw to se diagonals dictate the right towers and ne to sw diagonals dictate the left towers . there is one caveat : the _ empty _ tower , @xmath435 is now allowed . thus one has to determine from context whether a graph consisting of a single tower is of the form @xmath436 or @xmath437 . alternatively , we can picture the se corner as sitting directly above the ne corner of the lattice to form a half - diamond . thus proposition [ casebn ] is proven . on the next page , we give the lattice corresponding to @xmath438 . notice that there are six graphs in the northeast corner and four graphs in the southeast corner which are also graphs corresponding to positive roots and cluster variables for the @xmath439 case . in fact , if we decrease each label by @xmath88 and horizontally reflect the southeast corner , we can fit these two pieces together to obtain the @xmath440 lattice of graphs exactly . also comparing with the @xmath441 lattice we notice boundary behavior . for example the second entry of the third column is now @xmath442 instead of @xmath443 and the second to last element of the fourth column is @xmath444 instead of @xmath445 . we were able to analyze @xmath2 based on @xmath315 using a folding procedure . analogously we can analyze @xmath3 using @xmath447 and an _ unfolding _ procedure . we label the dynkin diagram for @xmath3 starting with @xmath10 and @xmath448 on the left , and label the rest in a line from @xmath88 to @xmath86 . indexing the rows and columns in the order @xmath449 , the corresponding exchange matrix is therefore @xmath450 we split the odd and even initial variables into the first two rows , in a zig - zagging pattern , just as before . we then mutate in the order @xmath10 , @xmath448 , @xmath84 , @xmath451 , @xmath433 , @xmath45 ( resp . @xmath86 ) if @xmath45 is odd ( resp . even ) to get the the third row , followed by mutation via @xmath88 then @xmath43 , @xmath452 , @xmath433 @xmath86 ( resp . @xmath45 ) if @xmath45 is odd ( resp . even ) to get the fourth row . the advantage of such an ordering is that the mutated exchange matrix , which we use to encode the binomial exchanges , is always the same , up to sign . we notice that the analogue of the diamond condition for this case is @xmath453 if @xmath454 with @xmath367 and @xmath455 on the western boundary . we let @xmath456 be @xmath457 where @xmath458 is the same tile as @xmath67 except with a different label . we also change tile @xmath309 so that it is still a hexagon , but has weights @xmath459 and @xmath326 going around clockwise from the top . following the arguments of lemmas @xmath451 , @xmath452 , @xmath460 , and @xmath461 result in the same graph theoretic interpretation and lattice structure . we use rule @xmath84 which is analogous to rule @xmath88 notice that when we apply rule @xmath10 to set of tiles @xmath456 , we get a set of graphs consisting of a base of @xmath309 or @xmath462 adjoining a tower of @xmath421 , as before . we enlarge the set of graphs by allowing a base of @xmath463 ( with or without an accompanying tower ) , and also allow _ both _ tile @xmath67 and tile @xmath458 to appear if and only if the lift of the graph to @xmath464 ( i.e. @xmath45 arbitrarily large ) is of the following three forms : where @xmath465 and @xmath337 both odd . let @xmath456 be defined as above and @xmath466 be the set of graphs constructed according to rules @xmath10 and @xmath84 . in particular , this construction will be quite analogous to that of @xmath467 . [ casedn ] the set @xmath466 is in bijection with the set of non - initial cluster variables for a coefficient - free cluster algebra of type @xmath3 such that the statement of theorem [ vargraph ] holds . + tiles for @xmath468 . as indicated , the proof follows from the exact same logic as lemmas @xmath451 through @xmath461 . the only caveat is that as a consequence of the proof , that @xmath372 will sometimes be a tower on base @xmath462 , and sometimes contain base @xmath463 . in particular , @xmath372 contains @xmath67 if and only if @xmath191 is odd , and so we get an alternating behavior . on the next page , we give the lattice for @xmath469 . we have the usual diamond condition for four entries in three consecutive rows and three consecutive columns , not including columns one . we encode column one by placing @xmath372 on top of @xmath470 , and we have the exchange relations ( [ dexch1 ] ) , ( [ dexch2 ] ) , and ( [ dexch3 ] ) . the case of @xmath94 is the only cluster algebra of exceptional finite type for which we have been able to extend our graph theoretic interpretation . we are able to do so since this case is analogous to @xmath471 . we use collection @xmath472 with tile @xmath67 as in the @xmath1 case , and tile @xmath309 is again a hexagon , but now has all three nontrivial weights being value @xmath325 . there are six possible graphs that correspond to the non - initial cluster variables . @xmath94 has dynkin diagram and exchange matrix @xmath474 given the previous sections , coefficient - free cluster algebras of type @xmath0 , @xmath1 , @xmath2 , @xmath3 , or @xmath94 have a combinatorial interpretation as a family of graphs such that the numerators of the cluster variables enumerate the weighted number of matchings and the denominators encode the occurrences of faces . thus theorem [ vargraph ] is true in all of these cases . the next step would be to extend theorem [ vargraph ] to include cluster algebras of type @xmath475 , @xmath476 , @xmath477 , and @xmath478 , and thus have the result for all cluster algebras of finite type . even though the dynkin diagrams for the @xmath479 s are simply laced , fitting cluster algebras of these three types into patterns analogous to those of the @xmath0 s and @xmath3 s has been notoriously hard . such difficulties have rose elsewhere such as in the original proof of positivity in @xcite , and also in recent models using @xmath480-paths on triangulated surfaces , for example in @xcite or @xcite among other work . additionally , in the work of schiffler and carroll - price for @xmath0 , the cluster algebra considered is specifically the ptolemy algebra , a cluster algebra _ with _ coefficients . in the @xmath480-paths model , the boundary of the polygon gives rise to @xmath481 additional coefficients which can be included in the exchange relations and cluster expansion formula . since the graphs we obtain in the above combinatorial interpretations are weighted so sparsely , perhaps a certain number of coefficients can be handled by the graph - model as well . in @xcite , an analogous interpretation is given for rank @xmath88 cluster algebras of affine type and unpublished work @xcite done as a part of reach , as described in @xcite , gives a graph theoretic interpretation for a totally cyclic rank @xmath84 cluster algebra . this totally cyclic rank @xmath84 cluster algebra corresponds to a triangulated surface of genus one with exactly one puncture ( i.e. interior marked point ) . such a cluster algebra has been studied geometrically including work of @xcite . perhaps these graph theoretical interpretations could be extended to other cluster algebras thus providing proofs of fomin and zelevinsky s positivity conjecture for even further cases . lastly , we note that all the examples discussed above are families of _ planar _ graphs associated to generators of cluster algebras . when expanding our scope to include more complicated cluster algebras , is the category of planar graphs too restrictive ? more specifically , why did we need the extra arcs in the @xmath1 , @xmath3 , @xmath94 , and affine @xmath482 cases ? perhaps it is an artifice of taking a higher dimensional object and projecting to two dimensions . the author would like to thank andrei zelevinsky for numerous helpful conversations including referring the author to @xcite where fibonacci polynomials appear . discussions with sergey fomin , jim propp , ralf schiffler , and hugh thomas have also been very useful . i wanted to especially thank hugh thomas and andrei zelevinsky for their comments on an earlier draft of this paper . i. assem , d. simson , and a. skowronski , _ elements of the representation theory of associated algebras . 1 . techniques of representation theory . london mathematical society student texts , * 65 . cambridge university press , cambridge , 2006 . *
in this paper we give a graph theoretic combinatorial interpretation for the cluster variables that arise in most cluster algebras of finite type . in particular , we provide a family of graphs such that a weighted enumeration of their perfect matchings encodes the numerator of the associated laurent polynomial while decompositions of the graphs correspond to the denominator . this complements recent work by schiffler and carroll - price for a cluster expansion formula for the @xmath0 case while providing a novel interpretation for the @xmath1 , @xmath2 , and @xmath3 cases .
to what extent ideal gas laws can be obtained from simple newtonian mechanics ? in this note , we present a simple one - dimensional simulation that suggest the answer is `` yes . '' . similar prior work includes the papers @xcite . they had a two dimensional model , in which two gases were seperated by a diabatic barrier . the way energy was communicated between the gases was via the barrier , in which after @xmath0 collisions , the particles which collided with the barrier in that time were reassigned equal energies , and momenta derived from these energies and their original momenta . in @xcite , this work was used to simulate the carnot cycle . this paper gives a slightly different model . in this paper , the barrier is a massless wall , which is pushed by the particles colliding into it , but the barrier keeps its vertical orientation so that only the horizontal positions of the gas particles matter in the dynamics of the barrier . an advantage to the approach given in this paper is that the motion is completely newtonian . for example , the motion is reversible , thus demonstrating that the second law of thermodynamics is not a hard and fast law , but rather is a trend generated by the pseudo - statistical nature of the deterministic dynamics . a thermally isolated container has a freely moving piston that splits the container into two parts , the left part and the right part . the piston is thermally conducting . the left part is filled with a monatomic gas a , and the right part of filled with a monatomic gas b. gas a has an atomic weight @xmath1 , and gas b has an atomic weight @xmath2 . initially , all the molecules have the same average velocity , picked independently from a symmetric random distribution . the piston is initially placed half way along the container . we place @xmath3 molecules of gas a randomly in the left half of the container , uniformly distributed , and we similarly place @xmath3 molecules of gas b in the right half of the container . the molecules obey newton s laws of motion , with elastic collisions . the objective of this note is to describe a numerical experiment which suggests that newton s laws of motion are all that are needed to predict the ideal gas laws , and that the collisions with the piston should be enough to transfer heat from the hotter gas to the colder gas . thus this note is in effect attempting to show that newtonian mechanics are completely sufficient to explain the second law of thermodynamics . molecules from the same gas are assumed sufficiently small so that collisions between them never occur . thus the molecules collide either with the walls of the container , or with the piston . if a molecule collides with a wall , it s velocity is simply reflected . however the piston has zero mass . thus if it is hit by a molecule , it simply travels with that molecule until it hits a molecule of the other gas . then the piston acts as a conduit for the two molecules to collide , conserving energy and momentum , as if they had the same @xmath4-component in space . ( -4,2)(4,2 ) ; ( 4,2)(4,-2 ) ; ( 4,-2)(-4,-2 ) ; ( -4,-2)(-4,2 ) ; ( -0.5,1.9)(-0.5,-1.9 ) ; ( -0.7,1.9)(-0.3,1.9 ) ; ( -0.7,-1.9)(-0.3,-1.9 ) ; ( -2.25,0 ) nodegas a ; ( 1.75,0 ) nodegas b ; from the ideal gas law , we have that each container , at least when in a state of quasi - equilibrium , should obey @xmath5 , where @xmath6 is the pressure , @xmath7 is the volume , @xmath8 is boltzmann s constant , @xmath3 is the number of molecules , and @xmath9 is the temperature . initially the temperature in the left hand part should be @xmath10 times cooler than the temperature in the right hand part . this is because the particles originally have the same average velocities , and hence the right hand side has total kinetic energy 100 times the kinetic energy of the left hand side , and kinetic energy is proportional to temperature . initially the pressure in the right hand side will be 100 times the left hand side , and so the piston should move quickly towards the left until the pressures are equalized . because this motion is quite fast , one might not expect conditions of quasi - equilibrium to be met , and so there will be some overshoot of this piston due to inertia of the gas molecules . next , because of the collisions between the molecules of the different gases , one might expect the kinetic energy to begin to spread evenly between the molecules in the left hand side and the right hand side . thus the piston should begin to move slowly towards the middle . to compute this numerically , we will completely neglect the vertical motions , and simply compute in one dimension . we let @xmath11 , and assign to each particle a one dimensional position and momentum . the only collisions that take place are between molecules and the left or right wall of the container , or between two molecules from the two different gases . a molecule from gas a will always be to the left of a molecule of gas b. molecules from the same gas will simply pass through each other . the computer program is as follows : 1 . assign to the @xmath3 particles of gas a positions uniformly picked from @xmath12 , and to the @xmath3 particles of gas b positions uniformly picked from @xmath13 $ ] . 2 . assign to all @xmath14 particles velocities uniformly picked from @xmath15 $ ] . ( the convention is that velocity is positive if the motion is to the right . ) [ loop ] compute the following collision times : 1 . the time for each particle to collide with the walls @xmath16 or @xmath17 ; 2 . the time for a particle of gas a to collide with a particle of gas b. 4 . let @xmath18 be the smallest of all these collision times . advance the positions of all the particles to time @xmath18 using their given velocities . if @xmath18 represents a collision of a particle with a wall , multiply the velocity of that particle by @xmath19 . if @xmath18 represents a collision of a particle of gas a with velocity @xmath20 , and a particle of gas b with velocity @xmath21 , then replace these velocities by @xmath22 and @xmath23 , where @xmath24 8 . go back to step . the program was carefully written so that floating point errors would not cause great error in the calculations . for example , when two particles collide , they might pass through each other by machine precision . but then when the collisions are recalculated , they will collide again in a very small amount of time . another place where care must be taken is when two collision events take place at exactly the same time . the collisions were stored in a database , implemented using c++ s standard library map . in this way , after a collision between two particles , or a particle and a wall , the only other collisions that need to be recalculated are those involving these two particles or particle and wall . the results , shown in figures [ plot1 ] , [ plot2 ] and [ plot3 ] , seem to bear out the prediction that newtonian mechanics is sufficient to obtain the ideal gas laws and the second law of thermodynamics . in figure [ plotxxx ] , we plot the moving averages over a range of 20 time units , showing better than expected agreement with charles law . the moving averages were calculated using the following formula . @xmath25 for an ideal gas , maxwell s distribution would predict that the @xmath26-componant of the velocities of the particles of either gas a or gas b should converge to a normal distribution . we decided to use the anderson - darling test described in @xcite . we assumed that both the mean and variance of the distributions were unknown . for each time @xmath18 , we calculated the modified statistic @xmath27 as described in table 1.3 in @xcite , and as case 4 in @xcite . this test assumes that the null - hypothesis is that the distribution is gaussian , and for a single application of the test , the hypothesis fails with 15% significance if @xmath28 , and fails with 1% significance if @xmath29 . in figure [ plotlr ] , we plot this statistic as a function of time . this is technically not a proper use of the statistic , as the statistic calculated at one particular time is not going to be statistically independent from the statistic calculated at another particular time . nevertheless , the graphs do strongly suggest that both gas a and gas b achieve a maxwellian distribution . as might be expected , the gas with the heavier molecules , gas b , takes longer to reach this distribution . statistic described in the text.,title="fig : " ] statistic described in the text.,title="fig : " ] another experiment that was performed was to run the numerics for 1000 time units , and then reverse the velocities , and then run the numerics backwards . it was found that in most cases that the data after 2000 time units matched the initial data to within about @xmath30 . however in one case the final data was completely different from the initial data . to help explain why it might be different just one time , the same experiment was performed many times , with only 100 particles and 100 time units . ( the 1000/1000 experiment takes about 3.5 hours . ) in these cases small random perturbations were made to the data at the same time as the velocities were reversed . by adjusting the random fluctuations to a size of about @xmath31 it was found that the initial data was recovered to within @xmath32 with about the same frequency as the final data being in complete disagreement with the initial data . in these experiments , a careful log of the order of the collisions was made , and the times when the data agreed closely , the order of collisions in the first 100 time units was almost exactly the reverse of the order of collisions in the second 100 time units ( the order differing with at most a few pairs of adjacent collisions reversing their order ) . when the final data greatly disagreed with the initial data , at some point in time the collisions became very different . these all suggest that the apparent statistical nature of equilibrium is not due to numerical floating point errors . it also suggests that if the final data has sensitive dependence on initial conditions , that this only happens when the time elapsed , or the number of particles , is very large . the effect of `` chaos '' is not primarily responsible for the statistical nature of the equilibrium . in this section , we derive the well known relationship between temperature and volume of an ideal gas when subjected to an adiabatic expansion . pressure is a more difficult quantity to understand in our context , so we will leave it out of our discussion . we have gas particles in a one - dimensional container , whose right hand side is able to move . adiabatic means that when the particles collide with the sides , we assume that the collisions are elastic , and the mass of the side walls are effectively infinite . the volume , @xmath33 , is proportional to the distance , @xmath34 , between the left hand side and the right hand side . the temperature is proportional to the average left - right kinetic energy of the particles , that is , if the @xmath3 particles have masses @xmath35 , and left - right - coordinate of velocity @xmath36 , for @xmath37 , then @xmath38 the particles themselves do not interact with each other . for this reason , without loss of generality , we only need to perform the calculation for one particle . the only thermodynamic assumption we shall make is that the energy in each particle is spread between the left - right kinetic energy and the other energies equally . let us denote by @xmath39 the ratio of the total energy to the left - right energy of each particle . thus if the particle is a three dimensional monotomic gas , then @xmath40 , and if the particle is a three dimensional diatomic gas , then @xmath41 ( because of the extra rotational kinetic energy ) . and if the particle is a three dimensional diatomic gas with a bond that is elastic , then @xmath42 ( two extra degrees of freedom for the rotational kinetic energy , one extra degree of freedom for the kinetic energy of the bond oscillating , and one extra degree of freedom for the potential energy of the bond oscillating ) . the purpose of this section is to derive the well known adiabatic thermodynamic relation @xmath43 the right hand side wall will move uniformly at a velocity @xmath44 , and the particle will bounce back and forth between the left and right wall . we assume that the @xmath36 are much larger than @xmath44 , and in the end we shall take the limit as @xmath45 . we also assume that the time for a back and forth bounce is much smaller than the time scale that @xmath44 might vary . if @xmath44 is allowed to vary too quickly , then we can create a kind of maxwell s daemon " in which we can impart any energy we like to the particle . the mass of the wall is taken to be @xmath46 , which we will take to be effectively infinite . we denote the velocity of the @xmath0th particle before and after it collides with the right wall are @xmath36 and @xmath47 respectively . the velocity of the wall after the collision is @xmath7 , and we will soon see that @xmath48 . ( 4,1)(4,-1 ) ; ( -4,-1)(-4,1 ) ; ( -3.1,0.5)node[above]@xmath49(4,0 ) ; ( 3.8,-0.05)node[below]@xmath50(-2.9,-0.5 ) ; ( 3.5,-1.5) node[below]@xmath44(4.5,-1.5 ) ; ( -4,1.5)node[above]@xmath34(4,1.5 ) ; let us consider one of the particles . conservation of momentum implies @xmath51 conservation of energy implies @xmath52 solving these equations we obtain @xmath53 now letting @xmath54 , we obtain @xmath55 the time @xmath56 required for the particle to bounce from the right hand wall , to the left hand wall , and back to the right hand wall is @xmath57 . in that time , the length of the wall changes @xmath58 , and the change in absolute value of the velocity of the particle is @xmath59 . then we obtain @xmath60 now let @xmath45 . then we obtain @xmath61 averaging over @xmath0 we obtain @xmath62 the authors feel that there are many extensions to these kinds of numerical simulations that could provide more illustrations of certain properties of ideal gases . for example , in the first section we could perform the experiment assuming that a constant force is applied to the piston , or we could assume the piston has non - zero mass . potentially there is a large number of variations that could be performed . we also feel that the scenario described in the first section could be a place to look for plausible theorems that could be rigorously proved . in particular , it seems to the authors that equilibrium is achieved rather quickly , and probably much more quickly than could be proved if we assumed certain ergodicity assumptions on the phase space of solutions . perhaps equilibrium theorems could be proved without ergodicity assumptions . and of course any equilibrium results must only show that equilibrium is achieved most of the time , because as our backwards in time calculations show , in this discrete setting the second law of thermodynamics is more of a trend than a law . adam galant , ryszard kutner , andrzej majerowski , heat transfer , newton s law of cooling and the law of entropy increase simulated by the real - time computer experiment in java , computational science iccs 2003 lecture notes in computer science volume 2657 , 2003 , pp 45 - 53 .
the purpose of this note is to see to what extent ideal gas laws can be obtained from simple newtonian mechanics , specifically elastic collisions . we present simple one - dimensional situations that seem to validate the laws . the first section describes a numerical simulation that demonstrates the second law of thermodynamics . the second section mathematically demonstrates the adiabatic law of expansion of ideal gases .
the _ spectrum _ @xmath0 of a finite group @xmath1 is the set of element orders of @xmath1 . finite groups having the same spectra are said to be _ isospectral_. recently , the following general result was obtained . suppose that @xmath3 is a finite simple group , and @xmath1 is a finite group with @xmath8 and @xmath9 . then @xmath1 is isomorphic to @xmath3 . the statement was conjectured by w. shi @xcite in 1987 , while its proof was completed in @xcite ( see the latter article for background information and complete list of references ) . it is worth mentioning that theorem a together with ( * ? ? ? * corollary 5.2 ) implies that a finite simple group and a finite group with the same burnside rings are isomorphic as well . what happens if we omit the condition @xmath9 in theorem a ? then @xmath1 is not necessary isomorphic to @xmath3 . for example , there are infinitely many groups with the same spectrum as the spectrum of the alternating permutation group of degree six @xcite . on the other hand , it turns out that for a bulk of finite nonabelian simple groups @xmath3 , a finite group isospectral to @xmath3 is isomorphic to @xmath3 or , at least , to a group @xmath1 with @xmath10 . investigations on this subject have 30-year history and resulted in more than a hundred papers of numerous authors . we do not intend to give a detailed review , rather prefer to draw an overall picture . in order to do that we formulate the following conjecture attributed to v.d . mazurov : for every finite nonabelian simple group @xmath3 , apart from a finite number of sporadic , alternating and exceptional groups and apart from several series of classical groups of small dimensions , if a finite group @xmath1 is isospectral to @xmath3 then @xmath1 is an almost simple group with socle isomorphic to @xmath3 . the conjecture was proved for sporadic groups @xcite , for alternating groups @xcite , and very recently for exceptional groups of lie type @xcite . here we deal with groups isospectral to finite simple classical groups . observe that there are a lot of results on this topic concerning groups in particular characteristics or dimensions . the peculiarity of our approach is that we sacrifice groups of small dimensions in order to obtain a general result ( cf . the theorems below with recent papers @xcite ) . throughout this paper we use single - letter names for simple classical groups , following @xcite , i.e. , for example , @xmath11 means @xmath12 , as well as the standard abbreviation @xmath13 , where @xmath14 , @xmath15 , and @xmath16 . [ t : l&u ] suppose that @xmath17 is a finite simple linear or unitary group and @xmath18 . then a finite group isospectral to @xmath3 is isomorphic to a group @xmath1 with @xmath10 . in particular , there are only finitely many pairwise non - isomorphic finite groups @xmath1 with @xmath8 . [ t : s&o ] suppose that @xmath3 is a finite simple symplectic or orthogonal group , where @xmath19 for @xmath20 , @xmath21 for @xmath22 , and @xmath23 for @xmath24 . if @xmath1 is a finite group with @xmath8 , then @xmath1 has a unique nonabelian composition factor @xmath6 , and one of the following holds : _ ( i ) _ @xmath25 _ ( ii ) _ @xmath20 and @xmath26 _ ( iii ) _ @xmath27 is even , @xmath22 , and @xmath28 in particular , there exist at most @xmath29 possibilities for @xmath6 for given @xmath3 . it seems very likely that the conclusion of theorem [ t : l&u ] is valid under the hypothesis of theorem [ t : s&o ] . in fact , as shown in the last section , theorems [ t : l&u ] and [ t : s&o ] are straightforward consequences of a series of previous results @xcite and the following theorem whose proof is the main goal of this paper . [ t : main ] let @xmath3 be a simple classical group over a finite field of characteristic @xmath4 , and @xmath1 be a finite group with @xmath8 . suppose that @xmath18 for @xmath17 , @xmath19 for @xmath20 , @xmath21 for @xmath22 , and @xmath23 for @xmath24 . then @xmath1 has a unique nonabelian composition factor @xmath6 , and @xmath6 is not isomorphic to a group of lie type over a field of characteristic distinct from @xmath4 . we strongly believe that the conclusion of theorem [ t : main ] remains true for all simple classical groups except the well - known examples of isomorphic groups in different characteristics such as @xmath30 ; three groups @xmath31 , @xmath32 , and @xmath33 , which are the only simple groups of lie type isospectral to some solvable groups ( * ? ? ? * corollary 1 ) ; and a few exotic cases such as @xmath34 @xcite . moreover , for many classical groups of small dimensions and specific characteristics the conclusion of the theorem has already been proved . in the article we concentrate on groups of sufficiently large dimensions in order to achieve a generic proof covering classical groups of all types in all characteristics . a final remark . the determination of properties of a group by means of its element orders is widely applied in computational group theory , especially in development of the so - called black - box algorithms , i.e. algorithms that do not exploit specific features of a group representation . we mention here just one of the numerous results on this subject , which is nearest to our main assertion . namely , w. kantor and . seress in @xcite proved that the characteristic of a finite simple group @xmath1 of lie type can be determined if three greatest element orders of @xmath1 are known ( it is additionally assumed that the characteristic of @xmath1 is odd ) . one may observe that theorem [ t : main ] says the same thing but only for classical groups of large dimensions and involving the whole spectrum of @xmath1 . however , we do not presuppose that @xmath1 is a simple group , and this is an important distinction between our approach and that of @xcite . on the other hand , in contrast to @xcite we do not propose here any practical implementation of our results . for nonzero integers @xmath35 , let @xmath36 and @xmath37 $ ] denote their greatest common divisor and least common multiple , respectively . given a nonzero integer @xmath27 , we put @xmath38 for the euler totient function of @xmath27 , @xmath39 for the set of prime divisors of @xmath27 , and if @xmath1 is a finite group then , as usual , @xmath40 stands for @xmath41 . if @xmath42 is a set of primes , then @xmath43 denotes the @xmath42-part of @xmath27 , that is , the largest divisor @xmath44 of @xmath27 with @xmath45 ; and @xmath46 denotes the @xmath47-part of @xmath27 , that is , the ratio @xmath48 . if @xmath27 is a nonzero integer and @xmath49 is an odd prime with @xmath50 , then @xmath51 denotes the multiplicative order of @xmath27 modulo @xmath49 . given an odd integer @xmath27 , we put @xmath52 if @xmath53 , and @xmath54 otherwise . fix an integer @xmath55 with @xmath56 . a prime @xmath49 is said to be a _ primitive prime divisor _ of @xmath57 if @xmath58 . we write @xmath59 to denote some primitive prime divisor of @xmath57 , if such a prime exists , and @xmath60 to denote the set of all such divisors . zsigmondy @xcite proved that primitive prime divisors exist for almost all pairs @xmath61 . [ l : zsigmondy ] let @xmath55 be an integer and @xmath56 . for every natural number @xmath62 the set @xmath60 is nonempty , except for the pairs @xmath63 . for @xmath64 the product of all primitive divisors of @xmath57 taken with multiplicities is denoted by @xmath65 . put @xmath66 . the number @xmath65 is said to be the _ greatest primitive divisor _ of @xmath57 . it follows from the definition that @xmath67 if @xmath68 . it is easy to check that @xmath69 if @xmath70 , and @xmath71 if @xmath72 , as well as @xmath73 if @xmath74 , and @xmath75 if @xmath76 . it follows from @xcite that for @xmath77 , @xmath78 where @xmath79 is the @xmath62th cyclotomic polynomial and @xmath49 is the largest prime dividing @xmath62 ; moreover , if @xmath80 does not divide @xmath81 then @xmath82 . note that for a divisor , the property of being primitive depends on the pair @xmath61 and is not determined by the number @xmath57 . for example , @xmath83 but @xmath84 , and @xmath85 but @xmath86 . [ l : kjpu0 ] suppose that @xmath55 and @xmath62 are integers with @xmath56 and @xmath87 , and @xmath4 is a prime . then @xmath88 divides @xmath89 . furthermore , if @xmath4 divides @xmath62 then @xmath90 . let @xmath49 be odd . then @xmath91 means that the multiplicative order of @xmath55 modulo @xmath49 equals @xmath92 . hence the order of @xmath93 modulo @xmath49 equals @xmath62 , that is @xmath94 . vice versa , if the order of @xmath93 modulo @xmath49 is equal to @xmath62 , and @xmath4 divides @xmath62 , then the order of @xmath55 modulo @xmath49 is equal to @xmath92 . thus , for odd @xmath88 and @xmath89 the assertion holds by the definition of a greatest primitive divisor . assume that @xmath95 . then @xmath55 is odd , @xmath96 , and @xmath97 . therefore , @xmath98 divides @xmath99 , and so divides @xmath100 . if @xmath101 and @xmath4 divides @xmath62 , then @xmath55 is odd , @xmath96 , and @xmath102 , so @xmath103 . [ l : kipma ] let @xmath55 and @xmath62 be integers with @xmath56 and @xmath87 . if @xmath62 is odd then @xmath104 , and if @xmath62 is a multiple of @xmath105 then @xmath106 . let @xmath62 be odd . by the definition of @xmath107 , we may assume that @xmath108 , so both @xmath109 and @xmath110 are odd . the order of @xmath55 modulo an odd prime @xmath49 is equal to @xmath111 if and only if the order of @xmath112 modulo @xmath49 is equal to @xmath62 . hence @xmath113 . it follows that @xmath104 because @xmath110 divides @xmath114 . the latter assertion follows from lemma [ l : kjpu0 ] . indeed , put @xmath115 and observe that @xmath116 . [ l : ki ] suppose that @xmath55 , @xmath62 , and @xmath117 are integers with @xmath56 , @xmath87 , and @xmath118 , @xmath49 is an odd prime such that @xmath119 and @xmath49 divides @xmath120 . the following hold : _ ( i ) _ if @xmath121 , then @xmath122 divides @xmath123 _ ( ii ) _ if @xmath124 , then @xmath122 divides @xmath125 _ ( iii ) _ if @xmath126 , then @xmath122 divides @xmath127 _ ( iv ) _ if @xmath128 , then @xmath122 divides @xmath129 _ ( v ) _ if @xmath130 , then @xmath122 divides @xmath131 _ ( vi ) _ if @xmath132 , then @xmath122 divides @xmath133 in particular , @xmath134 . moreover , @xmath135 in _ ( ii ) _ and _ ( v)_. observe that , by ( [ eq : ki ] ) , for @xmath136 and @xmath137 we have @xmath138 . it is easy to verify the following equalities : if @xmath121 , then @xmath139 for even @xmath55 , and @xmath140 for odd @xmath55 ; if @xmath124 , then @xmath141 ; if @xmath126 , then @xmath142 ; if @xmath143 , then @xmath144 if @xmath145 , then @xmath146 if @xmath147 , then @xmath148 . these equalities yield the lemma . [ l : estimknbyvarphi ] let @xmath55 and @xmath62 be integers , and @xmath14 . if @xmath149 , @xmath150 , and @xmath151 , then @xmath152 . we prove that @xmath153 first . let @xmath49 be the greatest prime divisor of @xmath62 and @xmath154 where @xmath155 . it follows that @xmath156 , and if @xmath81 is not a multiple of @xmath44 , then @xmath157 . if @xmath157 , then the desired inequality holds by ( * ? ? ? * lemma 7 ) , so we assume that @xmath49 divides @xmath158 and , in particular , @xmath44 divides @xmath81 . as observed in @xcite , the inequality @xmath159 where @xmath160 , holds true . since @xmath161 , we have @xmath162 let @xmath163 and @xmath164 . then @xmath165 and , therefore , @xmath166 by ( [ eq : estimkn ] ) . let @xmath167 and @xmath168 . then @xmath169 , @xmath170 . furthermore , @xmath44 divides @xmath171 . if @xmath172 , then @xmath173 , so we may assume that @xmath174 and @xmath175 . in this case the assertion follows because @xmath176 . let @xmath177 . then @xmath172 and @xmath178 . if @xmath179 then @xmath180 . therefore , by ( [ eq : estimkn ] ) , @xmath181 thus , @xmath182 , so @xmath183 and @xmath170 . since @xmath184 , the case @xmath169 is impossible by the hypothesis . if @xmath185 then @xmath186 . let , finally , @xmath187 . then @xmath188 , as required . thus , the inequality @xmath153 is proved . now we apply lemma [ l : kipma ] . if @xmath189 then @xmath106 . if @xmath190 then @xmath191 , so @xmath192 . finally , if @xmath62 is odd then @xmath193 , hence @xmath194 . the lemma is proved . define the following function on positive integers : @xmath195 [ l : kjubyu7 ] let @xmath196 be a prime power , @xmath14 , @xmath4 be an odd prime , and @xmath197 be a natural number with @xmath198 . then @xmath199 and @xmath200 . by lemma [ l : estimknbyvarphi ] , the assertion holds for sufficiently large @xmath197 . indeed , if @xmath201 , then @xmath202 @xmath203 where the latest inequality follows because @xmath204 . therefore , we may assume that @xmath205 and , in particular , @xmath206 . let @xmath207 stand for the set of @xmath197 , satisfying the hypothesis of the lemma and the inequality @xmath205 . by brute - force attack , we obtain that @xmath208 , @xmath209 , @xmath210 , @xmath211 , @xmath212 , @xmath213 , @xmath214 , @xmath215 , @xmath216 , @xmath217 , @xmath218 . if @xmath219 , then @xmath220 by ( [ eq : ki ] ) , so it is sufficient to prove the assertion for all @xmath221 , @xmath211 , @xmath213 , @xmath215 , @xmath216 , @xmath217 , @xmath218 . we use the following inequalities on @xmath222 for @xmath223 : @xmath224 @xmath225 @xmath226 @xmath227 @xmath228 @xmath229 @xmath230 by ( [ eq : ki ] ) , for @xmath231 , @xmath211 , @xmath213 , @xmath216 the equality @xmath232 holds , so @xmath199 by the above inequalities . for the other numbers from @xmath233 we have : @xmath234 @xmath235 @xmath236 thus , the first required inequality is proved . if @xmath4 divides @xmath197 then @xmath237 by lemma [ l : kjpu0 ] , so @xmath238 . therefore , we assume that @xmath4 does not divide @xmath197 . then @xmath239 . suppose that @xmath240 . then @xmath4 is the greatest prime divisor of @xmath241 , hence , using ( [ eq : ki ] ) , we obtain @xmath242 suppose that @xmath243 . in this case the greatest prime divisor of @xmath241 is at most @xmath217 . it follows that @xmath244 unless @xmath197 equals @xmath216 or @xmath213 . for @xmath245 , the greatest prime divisor of @xmath241 is equal to @xmath4 , and @xmath246 is not a multiple of @xmath197 , hence @xmath247 let @xmath248 . then @xmath249 , and the assertion follows from the inequalities : @xmath250 @xmath251 @xmath252 the lemma is proved . [ l : rpart ] let @xmath253 and @xmath254 be integers greater than @xmath255 , and @xmath14 . _ ( i ) _ if an odd prime @xmath49 divides @xmath256 , then @xmath257 . _ ( ii ) _ if an odd prime @xmath49 divides @xmath258 , then @xmath49 divides @xmath259 . _ ( iii ) _ if @xmath256 is a multiple of @xmath105 , then @xmath260 . see , for example , ( * ? ? ? * chapter ix , lemma 8.1 ) . let @xmath261 $ ] denote the integer part of a real number @xmath262 . [ l : eta ] let @xmath263 be positive integers , @xmath264 , and @xmath265 . if @xmath266 is even then @xmath267 $ ] . if @xmath266 is odd then @xmath268 $ ] . since the equation @xmath269 has two integer solutions for odd @xmath197 and one solution for even @xmath197 , the cardinality of @xmath270 is equal to the sum of @xmath55 and the quantity of odd numbers in the interval @xmath271 $ ] . if @xmath266 is even then this quantity is equal to @xmath272 $ ] , and if @xmath266 is odd then it equals @xmath273 $ ] . [ l : intervalwithprime ] if @xmath27 is a natural number and @xmath23 , then the interval @xmath274 contains a prime . if , in addition , @xmath275 , then the interval @xmath276 contains a prime . the first part is proved in @xcite . it is shown in @xcite that for every natural number @xmath277 , the interval @xmath278 $ ] contains a prime . since @xmath279 , the second part holds for all sufficiently large @xmath27 ( precisely , for @xmath280 ) . for smaller @xmath27 the assertion can be verified directly . the _ prime graph _ @xmath281 of a finite group @xmath1 is the nonoriented graph with the vertex set @xmath40 and two distinct vertices @xmath49 and @xmath282 are adjacent if and only if @xmath283 . the notion of prime graph was introduced by g.k . gruenberg and o. kegel ( for this reason it is also called the gruenberg kegel graph ) . they established that a finite group with disconnected prime graph is either frobenius or @xmath284-frobenius group , or has a unique nonabelian composition factor with disconnected prime graph . williams @xcite published this result and started the classification of finite simple groups with disconnected prime graph . the classification was completed by a.s . kondrat@xmath285ev @xcite . the full list of nonabelian simple groups with disconnected prime graph can be found , e.g. , in ( * ? ? ? * tables 1a-1c ) . recall that an independent set of vertices or a _ coclique _ of a graph @xmath286 is any subset of pairwise nonadjacent vertices of @xmath286 . we write @xmath287 to denote the independence number of @xmath286 , that is the greatest size of coclique in @xmath286 . for a group @xmath1 , put @xmath288 . by analogy , for each prime @xmath49 , define the @xmath49-independence number @xmath289 to be the greatest size of cocliques containing the vertex @xmath49 in @xmath281 . for convenience , we refer to a coclique containing @xmath49 as an @xmath290-coclique . in @xcite there was proved the following assertion which is , in some sense , a generalization of the gruenberg kegel theorem ( below we give the statement of this result from ( * ? ? ? * theorem 1 ) ) . [ l : structure ] let @xmath1 be a finite group with @xmath291 and @xmath292 . then the following hold : _ ( i ) _ there exists a nonabelian simple group @xmath6 such that @xmath293 , where @xmath294 is the maximal normal soluble subgroup of @xmath1 . _ ( ii ) _ for every coclique @xmath295 of @xmath281 of size at least @xmath29 , at most one prime of @xmath295 divides the product @xmath296 . in particular , @xmath297 . _ ( iii ) _ one of the following holds : _ ( a ) _ every prime @xmath298 nonadjacent to @xmath284 in @xmath281 does not divide the product @xmath299 in particular , @xmath300 _ ( b ) _ there exists a prime @xmath301 nonadjacent to @xmath284 in @xmath281 ; in which case @xmath302 , @xmath303 , and @xmath304 or @xmath305 for some odd @xmath253 . let @xmath3 and @xmath1 be isospectral finite groups . it follows by the definition of prime graph that @xmath306 . therefore , if @xmath3 satisfies the hypothesis of lemma [ l : structure ] , then so does @xmath1 . in @xcite , for every finite nonabelian simple group , an adjacency criterion of its prime graph is developed and all cocliques and @xmath307-cocliques of greatest size in this graph are found , as well as @xmath308-cocliques of greatest size for groups of lie type in characteristic @xmath4 . this information and williams s classification , imply the following assertion which is the first step toward a proof of main results of the present paper . [ p : uniques ] let @xmath3 be a finite simple group of lie type different from the groups @xmath31 , @xmath32 , and @xmath33 . if @xmath1 is a finite group isospectral to @xmath3 , then the following hold true for @xmath1 : _ ( i ) _ there exists a nonabelian simple group @xmath6 such that @xmath293 , where @xmath294 is the maximal normal soluble subgroup of @xmath1 . _ ( ii ) _ for every coclique @xmath295 of @xmath281 of size at least @xmath29 , at most one prime of @xmath295 divides the product @xmath296 . in particular , @xmath309 . _ ( iii ) _ every prime @xmath298 nonadjacent to @xmath284 in @xmath281 does not divide the product @xmath296 . in particular , @xmath310 if @xmath3 is a finite nonabelian simple group with connected prime graph and is different from an alternating group , then it satisfies the hypothesis of lemma [ l : structure ] by @xcite . if @xmath3 has a disconnected prime graph then the existence and uniqueness of a nonabelian composition factor @xmath6 follow from the gruenberg kegel theorem and @xcite . by ( * ? ? ? * propositions 2,3 ) , the inequality @xmath292 and the insolubility of @xmath1 imply that ( ii ) and ( iii ) of lemma [ l : structure ] hold . it remains to observe that the exceptional case ( b ) of assertion ( iii ) of lemma [ l : structure ] does not hold by ( * ? ? ? * theorem 2 ) . the proposition is proved . for a classical group @xmath3 , we put @xmath311 to denote its dimension if @xmath3 is a linear or unitary group , and its lie rank if @xmath3 is a symplectic or orthogonal group . observe that @xmath312 in theorems 13 in introduction . [ p : notexcept ] suppose that @xmath3 is a finite simple classical group , @xmath313 if @xmath3 is linear or unitary , and @xmath314 if @xmath3 is symplectic or orthogonal . suppose that @xmath1 is a finite group isospectral to @xmath3 , and @xmath6 is a unique nonabelian composition factor of @xmath1 . if @xmath6 is a group of lie type , then @xmath6 is a classical group , @xmath315 if @xmath6 is linear or unitary , and @xmath316 if @xmath6 is symplectic or orthogonal . applying ( * ? ? ? * tables 2,3 ) , it is easy to obtain that @xmath317 provided the hypothesis of the proposition . by proposition [ p : uniques](ii ) , we have @xmath318 . on the other hand , it follows from ( * ? ? ? * table 4 ) that @xmath319 for every simple exceptional group @xmath2 of lie type . thus , @xmath6 is a classical group . the required inequalities on @xmath320 hold by ( * ? ? ? * tables 2,3 ) . the proposition is proved . given a classical group @xmath3 over a field of order @xmath253 , put @xmath321 [ l : indepofcrit ] let @xmath3 be a simple classical group over a field of order @xmath253 and characteristic @xmath4 . suppose that @xmath49 and @xmath282 are distinct primes , @xmath322 , @xmath323 , and @xmath324 . _ ( i ) _ if @xmath325 , then @xmath326 for every distinct odd primes @xmath327 and @xmath328 . _ ( ii ) _ if @xmath329 , then @xmath330 for every odd prime @xmath331 . it follows from @xcite . thus , for two distinct primes @xmath332 , where @xmath333 , the answer to the question whether they are adjacent in @xmath334 depends only on @xmath335 , if @xmath336 , and @xmath335 , @xmath337 , if @xmath338 . in @xcite several functions of natural argument are used to formulate an adjacency criterion , one of them is @xmath339 defined in ( [ eq : eta ] ) , and two others we define here . @xmath340 for @xmath14 , put @xmath341 it is an easy observation that @xmath342 is a bijection and @xmath343 is an identity . for linear and unitary groups , we exploit also a reformulation of an adjacency criterion ( see ( * ? ? ? * lemmas 2.12.3 ) ) , if it is more convenient for our goals than an initial formulation from @xcite which used the function @xmath342 . this reformulation is based on the equality @xmath344 , which follows from lemma [ l : kipma ] and the definition of @xmath342 . now we introduce a new function in order to unify further arguments . namely , given a simple classical group @xmath3 over a field of order @xmath253 and a prime @xmath49 coprime to @xmath253 , we put @xmath345 it follows that @xmath346 observe that @xmath347 in the case of @xmath348 . [ l : fermat ] let @xmath3 be a simple classical group over a field of order @xmath253 and characteristic @xmath4 . if @xmath49 is an odd prime from @xmath349 then @xmath350 divides @xmath81 , and if @xmath3 is a symplectic or orthogonal group then @xmath351 divides @xmath81 . if @xmath3 is not unitary , then @xmath350 divides @xmath335 , which divides @xmath81 by fermat s little theorem . if @xmath3 is unitary , then @xmath350 does not divide @xmath335 only when @xmath335 is odd . but then @xmath352 and @xmath350 divides @xmath81 because @xmath49 is odd . let @xmath3 be a symplectic or orthogonal group . if @xmath350 is even then @xmath353 , and if not , then @xmath350 divides @xmath354 . [ l : auxvarphi ] let @xmath3 be a simple classical group over a field of order @xmath253 and characteristic @xmath4 , and let @xmath355 . _ ( i ) _ if @xmath356 , then @xmath357 . _ ( ii ) _ if @xmath49 and @xmath282 are distinct primes from @xmath349 with @xmath358 and @xmath359 , then @xmath49 and @xmath282 are adjacent in @xmath334 . _ _ if @xmath49 and @xmath282 are distinct primes from @xmath349 with @xmath360 and @xmath361 , then @xmath49 and @xmath282 are adjacent in @xmath334 if and only if @xmath362 . _ ( iv ) _ if @xmath49 and @xmath282 are distinct primes from @xmath349 and @xmath362 , then @xmath49 and @xmath282 are adjacent in @xmath334 . it follows from @xcite . let @xmath3 be a simple classical group over a field of order @xmath253 and characteristic @xmath4 . for @xmath363 , set @xmath364 . if @xmath365 then , by @xcite , every coclique @xmath295 of greatest size in @xmath334 does not contain @xmath4 , so the set @xmath366 is well - defined for @xmath295 . define @xmath367 as the union of sets @xmath366 , and @xmath368 as the intersection of these sets , where @xmath295 runs over all cocliques of greatest size in @xmath334 . the next lemma is a particular case of the main theorem of @xcite . [ l : tl ] let @xmath3 be a simple classical group over a field of order @xmath253 and characteristic @xmath4 , and let @xmath365 . let @xmath295 be a coclique of greatest size in @xmath334 . if @xmath369 then @xmath370 . if @xmath371 then @xmath372 for some @xmath373 . in particular , @xmath374 . the sets @xmath368 , @xmath375 and numbers @xmath376 are listed in table _ [ tab : tl]_. see ( * ? ? ? * tables 2,3 ) . .cocliques of greatest size [ cols="^,<,^,^,^",options="header " , ] let @xmath197 and @xmath44 be natural numbers and @xmath377 . observe that @xmath378 is an odd integer if and only if @xmath379 . [ l : kdividesj ] let @xmath380 . suppose that @xmath381 are different primes such that @xmath382 , @xmath383 , and @xmath384 , and suppose that @xmath385 and @xmath386 . if @xmath387 , then neither @xmath388 nor @xmath389 is divisible by @xmath390 . if @xmath6 is symplectic or orthogonal , then neither @xmath62 nor @xmath197 is congruent to @xmath44 modulo @xmath391 . in particular , if @xmath392 and @xmath393 , then @xmath388 is not divisible by @xmath390 when @xmath387 , and @xmath197 is not congruent to @xmath44 modulo @xmath391 when @xmath6 is symplectic or orthogonal . first , suppose that @xmath394 . to avoid unwieldy notation , we assume that @xmath6 is a linear group , that is @xmath342 is the identity function . the proof for unitary groups is analogous , we have only to use the function @xmath395 instead . recall that by lemma [ l : pdoesnotdivideum1 ] , the number @xmath4 does not divide @xmath396 , that is we have @xmath397 . it follows from the equality @xmath382 that @xmath6 has a semisimple element of order @xmath398 . now lemma [ l : criterion ] implies that @xmath6 has an element of order @xmath399/d$ ] , where @xmath400 is some divisor of @xmath396 . if @xmath44 divides either of @xmath197 and @xmath62 , then @xmath401 divides @xmath399/d$ ] , which contradicts the fact that @xmath402 . now let @xmath6 be a symplectic or orthogonal group . recall that @xmath403 . the equality @xmath382 implies that @xmath6 has a semisimple element of order @xmath398 . by lemma [ l : criterion ] , @xmath6 has an element of order @xmath404/d$ ] , where @xmath400 divides @xmath105 . assume that one of the numbers @xmath197 and @xmath62 , say @xmath197 , is congruent to @xmath44 modulo @xmath391 . then @xmath405 divides @xmath406 , and hence @xmath401 divides @xmath407/d$ ] , contrary to the hypothesis . let @xmath392 and let @xmath393 . by lemma [ l : rsnotprs ] , there is a number @xmath266 such that @xmath408 is large with respect to @xmath3 , @xmath409 lies in @xmath410 , and @xmath411 does not . writing @xmath412 , @xmath413 and observing that @xmath382 , we see that the assertion of the lemma holds for @xmath414 . [ l : lnotinx ] if @xmath415 , then the conclusion of theorem _ [ t : main ] _ holds . . then lemma [ l : sinyi ] implies that @xmath6 also lies in @xmath416 . suppose that @xmath417 . since @xmath415 , we have @xmath418 , and hence @xmath419 ( see table [ tab : kll ] ) . one of the numbers @xmath420 , @xmath421 , @xmath422 is a multiple of @xmath29 , and we denote this number by @xmath423 . by lemma [ l : tmaless2t3 ] , we have @xmath424 , and so @xmath425 . therefore , there is @xmath392 such that @xmath426 . then @xmath427 due to lemma [ l : transfer ] . by lemma [ l : tpll ] , the function @xmath428 is injective and @xmath429 , and hence @xmath430 . since @xmath431 , we see that @xmath432 is divisible by @xmath433 , which is a contradiction by lemma [ l : kdividesj ] . let @xmath6 be an orthogonal group lying in @xmath416 . then @xmath434 . by lemmas [ l : tplop ] and [ l : tplom ] , the set @xmath435 contains @xmath423 such that @xmath436 and @xmath437 . since @xmath438 , it follows that @xmath439 . thus there is @xmath392 such that @xmath426 . then @xmath440 , and hence @xmath441 , which is not the case by lemma [ l : kdividesj ] . [ l : sisnotx1 ] if @xmath442 , then the conclusion of theorem _ [ t : main ] _ holds . let @xmath380 . suppose that there is @xmath439 such that @xmath390 divides @xmath423 . as in the previous lemma , there is @xmath392 such that @xmath440 , and by lemma [ l : tpll ] , it follows that @xmath443 . hence @xmath432 is divisible by @xmath390 , which is a contradiction due to lemma [ l : kdividesj ] . thus it suffices to show that there is @xmath444 that is a multiple of @xmath390 . by the previous result , we may assume that @xmath445 . in this case , an additional difficulty is that the members of @xmath446 are not consecutive integers . nevertheless , for each group @xmath3 we will prove that @xmath446 contains a number with the desired property . let @xmath447 and @xmath27 be even . then @xmath448 . since @xmath449 , and therefore , the characteristic @xmath4 of @xmath3 does not divide @xmath450 by lemma [ l : pdoesnotdivideum1 ] , it follows that @xmath451 , and hence @xmath452 . if @xmath453 , then @xmath454 , and if @xmath455 , then @xmath456 . in each case , @xmath446 contains two consequent numbers , and one of them is a multiple of @xmath284 , as required . let @xmath24 and @xmath27 be even . then @xmath457 , whence @xmath458 . now the proof is somewhat similar to the proof in the previous paragraph if we take into account the following observation : @xmath446 contains the numbers @xmath459 when @xmath453 and @xmath459 when @xmath455 and @xmath460 , and the set of the residues of these numbers module 4 is all of @xmath461 ; furthermore , if @xmath462 , then @xmath463 is a multiple of @xmath105 and lies in @xmath376 , and if @xmath464 , then so does @xmath465 . the remaining cases are @xmath466 , @xmath467 with @xmath27 odd and @xmath22 with @xmath27 even , where we have @xmath418 , whence @xmath419 and @xmath433 . if @xmath466 , @xmath467 and @xmath468 , then @xmath469 . thus @xmath421 lies in @xmath446 and is a multiple of @xmath29 . if either @xmath466 , @xmath467 with @xmath470 or @xmath22 with @xmath453 , then @xmath471 , and @xmath422 lies in @xmath446 and is a multiple of @xmath29 . if @xmath22 and @xmath455 , then @xmath472 , and @xmath420 is the desired number . now we may assume that @xmath445 and @xmath473 , and in particular , both @xmath3 and @xmath6 are symplectic or orthogonal groups . moreover , by lemma [ l : sinyi ] , each @xmath474 satisfies the following condition : if @xmath475 , then @xmath476 . also observe that @xmath477 $ ] is an injective function , and hence @xmath478 is uniquely determined by @xmath479 . furthermore , if @xmath480 , then @xmath62 is uniquely determined by @xmath478 , and so by @xmath479 as well . [ l : tplneq3 ] if @xmath418 , then the conclusion of theorem _ [ t : main ] _ holds . let @xmath418 . then either @xmath466 , @xmath467 with @xmath27 odd or @xmath22 with @xmath27 even . furthermore , since @xmath6 is symplectic or orthogonal , we have @xmath434 . also note that if @xmath481 , @xmath482 , then @xmath483 ; if @xmath484 , then @xmath485 ; and if @xmath486 , then @xmath487 ( see table [ tab : kll ] ) . let @xmath466 , @xmath467 and @xmath468 . suppose that @xmath481 or @xmath488 . since @xmath489 , it follows that @xmath490 . now the congruence @xmath483 yields @xmath491 . but then @xmath492 , which is not the case because @xmath493 . if @xmath484 , then for every @xmath254 , we have that @xmath494 is not congruent to @xmath284 modulo @xmath29 . on the other hand , @xmath495 , and hence @xmath496 . similarly , @xmath497 , since otherwise @xmath487 and @xmath498 . let @xmath466 , @xmath467 and @xmath470 . suppose that @xmath481 or @xmath488 . the equality @xmath489 yields @xmath499 . both @xmath500 and @xmath501 are divisible by @xmath105 , therefore , one of them is congruent to @xmath105 modulo @xmath502 . denote this number by @xmath55 . since @xmath503 and @xmath504 , we have @xmath505 , whence @xmath506 . thus @xmath507 . as @xmath428 is invertible on the set of multiples of @xmath105 and @xmath6 has the same type as @xmath3 , it follows that @xmath508 is equal to @xmath55 , and so it is congruent to @xmath105 modulo @xmath502 . applying lemma [ l : kdividesj ] , we derive a contradiction . thus @xmath6 is an orthogonal group of even dimension . if @xmath499 , then @xmath509 . by lemmas [ l : tpls][l : tplom ] , both @xmath510 and @xmath511 are equal to @xmath421 , so @xmath512 . similarly , @xmath513 . repeating the previous argument we derive a contradiction by lemma [ l : kdividesj ] . if @xmath514 , then the equality @xmath489 implies that the only remaining possibility is @xmath515 . then @xmath492 , which is not the case because @xmath516 . let @xmath22 and @xmath455 . if @xmath517 , then @xmath489 yields @xmath518 , whence @xmath519 , which is a contradiction . if @xmath486 , then @xmath520 , and this is also a contradiction since @xmath254 must be congruent to @xmath29 modulo @xmath105 . if @xmath484 , then @xmath521 . since @xmath254 can not be congruent to @xmath255 modulo @xmath105 , it follows that @xmath522 . thus @xmath6 has the same type and the same lie rank @xmath27 as @xmath3 . both @xmath523 and @xmath524 are multiples of @xmath105 , so one of them is congruent to 4 modulo 8 , and we denote this number by @xmath55 . as @xmath525 , we have @xmath21 and @xmath526 . therefore @xmath392 , and applying lemma [ l : kdividesj ] , we derive a contradiction . let @xmath22 and @xmath453 . the equality @xmath489 leaves us with two possibilities : either @xmath527 or @xmath499 and @xmath528 . if @xmath6 is a symplectic group or an orthogonal group of odd dimension , then @xmath529 , while @xmath530 . suppose that @xmath528 , that is @xmath6 has the same type and the same lie rank @xmath27 as @xmath3 . since @xmath531 , @xmath532 are multiples of @xmath105 , we can choose a number @xmath55 of them with @xmath533 . furthermore , @xmath534 because @xmath525 and so @xmath21 . thus @xmath392 , and applying lemma [ l : kdividesj ] , we derive a contradiction . the remaining case for @xmath22 with @xmath453 , is @xmath535 , and this case requires the most effort . let @xmath536 and @xmath537 . similarly to the proof of lemma [ l : rsnotprs ] we conclude that @xmath538 and @xmath539 have the following properties : for @xmath540 , the number @xmath541 is large with respect to @xmath3 , @xmath542 , and @xmath543 . since @xmath453 , we conclude that @xmath544 has the same properties . now @xmath545 constitute a coclique in @xmath334 because for any @xmath546 we have @xmath547 and @xmath548 whenever @xmath68 . therefore these numbers are large with respect to @xmath6 and constitute a coclique in @xmath549 . since @xmath428 is injective and @xmath550 , the value of @xmath551 depends only on @xmath337 . if @xmath552 , then @xmath553 , and if @xmath554 , then @xmath555 . in both cases , @xmath556 is odd . let @xmath557 and write @xmath386 . as @xmath558 , we have @xmath559 . if @xmath560 , then since @xmath254 is odd , it follows that @xmath62 is even . the condition @xmath558 implies that @xmath6 has an element of order @xmath407/2 $ ] , which is not the case because @xmath62 and @xmath197 have opposite parity when @xmath484 , and the same parity when @xmath486 . thus @xmath561 . if @xmath562 , then @xmath6 has an element of order @xmath563 $ ] , which is divisible by @xmath401 , contrary to the fact that @xmath402 . let @xmath564 . then @xmath62 is even , and hence @xmath6 has an element of order @xmath565 $ ] , which is divisible by @xmath401 . finally , let @xmath566 . but then @xmath567 has only two different possible values , and this is a contradiction since the numbers @xmath568 with @xmath569 must be distinct . [ l : tplneq2 ] if @xmath448 , then the conclusion of theorem _ [ t : main ] _ holds . if @xmath448 , then @xmath570 and @xmath27 is even . since @xmath6 is symplectic or orthogonal , we have @xmath571 . . then @xmath530 . the equality @xmath489 implies that @xmath572 . let @xmath573 and @xmath574 . observe that @xmath575 is odd . as in the proof of lemma [ l : rsnotprs ] , we write @xmath576 . then @xmath577 and @xmath539 have the following properties : for @xmath540 , the number @xmath541 is large with respect to @xmath3 , @xmath542 , and @xmath543 . furthermore , @xmath578 and @xmath579 are not adjacent in @xmath334 . since @xmath580 and @xmath6 has the same type and the same lie rank as @xmath3 , it follows that @xmath581 . let @xmath582 and write @xmath386 . as @xmath558 , we have @xmath583 . if @xmath584 , then @xmath6 contains elements of orders @xmath585 $ ] and @xmath586 $ ] , one of which must be divisible by @xmath401 because @xmath4 divides either @xmath396 or @xmath587 . but then @xmath588 , a contradiction . thus @xmath589 . since @xmath55 is odd , there are two possible values for @xmath567 , and these are @xmath55 and @xmath590 . as @xmath591 , we have @xmath592 . if @xmath593 , then @xmath594 , and if @xmath595 , then @xmath596 . we denote by @xmath49 the number in the set @xmath597 for which @xmath598 . now @xmath282 and @xmath49 satisfy the hypothesis of lemma [ l : kdividesj ] , which is a contradiction because @xmath599 . now suppose that @xmath453 . the conditions @xmath489 and @xmath516 leave us with two further possibilities : either @xmath572 or @xmath600 with @xmath491 . in the first case , we use the argument of the previous paragraph with @xmath601 and @xmath602 . now assume that @xmath603 and write @xmath604 , @xmath605 , @xmath606 , @xmath577 , and @xmath539 . since @xmath607 is odd , it follows that @xmath578 and @xmath579 satisfy the conclusion of lemma [ l : rsnotprs ] , that is @xmath578 and @xmath579 are not adjacent in @xmath334 , they are large with respect to @xmath3 , and for @xmath540 the number @xmath608 belongs to @xmath410 , but @xmath609 does not . let @xmath582 , @xmath386 , and @xmath385 . then the equality @xmath610 yields @xmath611 when @xmath484 , and @xmath612 when @xmath486 . as @xmath558 , it follows that @xmath559 . if @xmath613 , then @xmath614 contains @xmath615 $ ] which is divisible by the odd part of @xmath616 , and so by @xmath401 , which is a contradiction . thus since @xmath617 , the number @xmath618 is even , and therefore there is only one possible value for @xmath567 , namely , @xmath619 . but then @xmath620 , a contradiction . [ l : tplneq4 ] if @xmath457 , then the conclusion of theorem _ [ t : main ] _ holds . note that the equality @xmath457 holds if and only if @xmath24 and @xmath27 is even . suppose that @xmath453 . if @xmath481 or @xmath488 , then @xmath489 yields @xmath499 . but then @xmath529 and @xmath493 , a contradiction . if @xmath486 , then @xmath490 . suppose that @xmath499 . then @xmath434 . since @xmath621 and @xmath622 lie in @xmath623 and are multiples of @xmath105 , it follows that one of the numbers @xmath624 and @xmath625 is congruent to @xmath105 modulo @xmath502 . but this contradicts lemma [ l : kdividesj ] . if @xmath484 , then @xmath626 . let @xmath627 . then @xmath628 , and , on the other hand , it follows from table [ tab : kll ] that @xmath629 . thus @xmath630 , and hence @xmath631 , @xmath632 for some integer @xmath423 . writing @xmath633 and noting that @xmath634 is even , we conclude that @xmath635=3c+3 $ ] by lemma [ l : eta ] . there is a coclique @xmath295 of size @xmath636 in @xmath334 such that @xmath637 for every @xmath638 . let @xmath638 . the prime @xmath49 is large with respect to @xmath3 , and so @xmath49 is large with respect to @xmath3 too , whence @xmath639 ( see table [ tab : tl ] ) . furthermore , there is @xmath640 such that @xmath641 . since @xmath642 and @xmath643 , we have @xmath644 . thus writing @xmath645 , we see that @xmath646 for every @xmath638 . by lemma [ l : eta ] , the size of @xmath647 is equal to @xmath648=3c+3=|x|$ ] , and therefore @xmath649 . it follows that @xmath295 contains @xmath49 and @xmath650 such that @xmath651 and @xmath652 . but then @xmath653 , a contradiction . it remains to consider the situation where @xmath600 , @xmath491 , and either @xmath434 or @xmath654 and @xmath655 . first , assume that @xmath654 and @xmath656 . observe that the congruence and the condition @xmath657 imply that @xmath658 . in particular , we have @xmath659 , and so @xmath660 with @xmath661 , @xmath662 , and @xmath663 . if @xmath484 , then it follows that the numbers @xmath664 , @xmath665 , and @xmath666 are congruent to @xmath284 modulo @xmath105 . moreover , they have different residues modulo @xmath29 , and hence one of them is congruent to @xmath667 modulo @xmath668 . we denote this number by @xmath197 and take @xmath55 to be such that @xmath441 . applying lemma [ l : kdividesj ] to @xmath392 , we conclude that @xmath669 , a contradiction . if @xmath486 , then the numbers @xmath670 , @xmath671 , and @xmath672 are odd and one of them is a multiple of @xmath29 . since @xmath673 , we again derive a contradiction by lemma [ l : kdividesj ] . thus we may assume that @xmath434 . let @xmath674 , @xmath675 , @xmath676 , @xmath677 , and @xmath678 . it follows that @xmath679 , the primes @xmath680 are large with respect to @xmath3 and for @xmath681 , we have @xmath682 , but @xmath683 . furthermore , @xmath684 . therefore , @xmath685 is equal to @xmath686 when @xmath484 and to @xmath687 when @xmath486 . let @xmath688 and @xmath386 . as for @xmath681 , we have @xmath689 , it follows that @xmath690 . if @xmath691 , then @xmath478 is even , and so @xmath692 . but then @xmath693 and either @xmath694 and @xmath484 or @xmath695 and @xmath486 , and by adjacency criterion , @xmath49 and @xmath696 are not adjacent in @xmath6 for @xmath681 , which is not the case . thus @xmath697 . if @xmath698 , then @xmath699 has an element of order @xmath700 $ ] and @xmath701 has an element of order @xmath702 $ ] . this is a contradiction because @xmath703 divides both @xmath55 and @xmath266 for @xmath681 . if @xmath704 , then it is easy to construct an element of required order . so @xmath705 and there are only two possible values for @xmath567 . this is impossible because @xmath706 is a coclique in @xmath334 . now suppose that @xmath455 . the group @xmath6 can not be isomorphic to @xmath699 since otherwise @xmath376 is congruent to @xmath284 modulo @xmath29 and @xmath494 does not . let @xmath707 . then @xmath489 yields @xmath708 . if @xmath499 , then @xmath492 , and if @xmath709 , then @xmath710 , which contradicts to the fact that @xmath516 . thus @xmath486 , and it follows from @xmath489 that @xmath499 . then @xmath628 , and so @xmath434 . assume that @xmath711 , that is @xmath712 . then @xmath713 , and hence @xmath714 with one of them being a multiple of @xmath105 but not of @xmath502 . now lemma [ l : kdividesj ] gives the desired contradiction . finally , assume that @xmath715 and @xmath716 . let @xmath717 , @xmath718 , @xmath719 , @xmath720 , and @xmath721 . then @xmath722 , @xmath706 is a coclique in @xmath334 and it consists of numbers large with respect to @xmath3 , and for @xmath681 , we have @xmath682 , but @xmath683 . writing @xmath723 , @xmath724 and using lemma [ l : transfer ] , we obtain @xmath725 . choose @xmath688 and take @xmath386 . since for @xmath681 , we have @xmath689 , it follows that @xmath726 . it is easy to see that if @xmath727 , then @xmath614 contains a number divisible by @xmath703 for each @xmath681 , which is impossible . so we assume that @xmath728 . if @xmath729 , then @xmath730 , and applying lemma [ l : kdividesj ] either to the pair @xmath731 or to the pair @xmath732 for some @xmath733 , we have the desired contradiction . let @xmath734 and @xmath735 . if @xmath736 , then since @xmath6 has no elements of order @xmath737/d$ ] with @xmath738 and does have element of order @xmath739 $ ] , it follows that @xmath740 . similarly , if @xmath741 then @xmath742 . if @xmath743 , then @xmath386 can take at most two different values , but there must be at least three of them as @xmath706 is a coclique in @xmath334 . and if @xmath744 , then @xmath745 , and this contradiction completes the proof . if @xmath17 with @xmath18 , @xmath447 with @xmath746 , @xmath24 with @xmath23 , or @xmath22 with @xmath21 , then @xmath525 . we eliminated all possibilities for the group @xmath6 under the assumption that @xmath525 , and so the proof of theorem [ t : main ] is complete . although the final part of the proof is technically complicated , its idea is transparent and based on an application of the well - known pigeonhole principle . here pigeons are from @xmath747 and holes are from @xmath748 . the number of pigeons and number of holes are almost the same and , what is sufficient , the difference between them is a small constant which does not depend on @xmath312 . however , there are prohibited holes , and the number of those holes increases when so does @xmath27 . it provides a contradiction that becomes evident with the growth of @xmath27 . as written in introduction , theorems [ t : l&u ] and [ t : s&o ] follow from theorem [ t : main ] and a series of previously obtained results . the next assertion summarizes the main results of @xcite . [ l : othercases ] let @xmath3 be a finite simple classical group over a field of characteristic @xmath4 , and @xmath749 . suppose that @xmath1 is a finite group with @xmath8 , and @xmath6 is a unique nonabelian composition factor of @xmath1 . then one of the following holds : _ ( 1 ) _ @xmath750 _ ( 2 ) _ @xmath751 , where @xmath752 , and @xmath753 _ ( 3 ) _ @xmath754 and @xmath755 _ ( 4 ) _ @xmath756 , @xmath20 and @xmath26 _ ( 5 ) _ @xmath757 is even , @xmath22 and @xmath758 _ ( 6 ) _ @xmath6 is a group of lie type over a field of characteristic distinct to @xmath4 . this assertion and theorem [ t : main ] yield theorem [ t : s&o ] immediately . a _ cover _ of a finite group @xmath1 is a finite group having @xmath1 as a factor group . a cover is proper if the corresponding factor group is proper . if @xmath759 for every proper cover @xmath2 of @xmath1 , then @xmath1 is said to be _ recognizable by spectrum among covers_. it follows from ( * ? ? ? * lemma 9 ) that a finite group @xmath1 is recognizable by spectrum among covers if and only if @xmath759 for every splitting extension @xmath760 , where @xmath761 is an absolutely irreducible finite - dimensional @xmath1-module over a finite field of characteristic @xmath49 . for every simple linear and unitary group @xmath17 with @xmath762 and every @xmath3-module @xmath761 , if the characteristic @xmath49 of @xmath761 coincides with the characteristic @xmath4 of @xmath3 , then it is proved that @xmath763 @xcite . if @xmath761 is defined over the field of characteristic distinct to @xmath4 and @xmath764 , then any extension @xmath765 contains an element whose order does not belong to @xmath410 ( * ? ? ? * lemma 11 ) . for other classical groups the following general result was recently obtained . [ l : grcov ] let @xmath3 be one of the simple groups @xmath766 , where @xmath767 , @xmath768 , where @xmath769 , @xmath467 , where @xmath770 , and @xmath771 , where @xmath767 . suppose that @xmath761 is a finite - dimensional @xmath3-module over a field of characteristic @xmath49 prime to @xmath253 . then either @xmath772 or @xmath773 and @xmath177 . if @xmath773 , then there is a @xmath774-dimensional @xmath3-module @xmath761 over a field of characteristic @xmath29 such that @xmath775 . if @xmath17 , @xmath18 , and @xmath1 is a finite group isospectral to @xmath3 , then theorem [ t : main ] and lemma [ l : othercases ] provide @xmath776 , where @xmath294 is the soluble radical of @xmath1 . an application of the aforementioned results on covers of linear and unitary groups completes the proof of theorem [ t : l&u ] . the author is grateful to m.a . grechkoseeva for her reading earlier drafts of this article . her valuable comments and helpful suggestions have enabled us to improve the text significantly . l. di martino , a. e. zalesskii , minimum polynomials and lower bounds for eigenvalues multiplicities of prime - power order elements in representations of quasi - simple groups , j. algebra 243 ( 2001 ) 228263 , corrigendum , j. algebra 296 ( 2006 ) , 249252 . e. i. khukhro , v. d. mazurov , automorphisms with centralizers of small rank , in : groups st . andrews 2005 , vol . 340 of london math . lecture note ser . , cambridge univ . press , cambridge , 2007 , pp .
the _ spectrum _ @xmath0 of a finite group @xmath1 is the set of element orders of @xmath1 . finite groups @xmath1 and @xmath2 are _ isospectral _ if their spectra coincide . suppose that @xmath3 is a simple classical group of sufficiently large dimension ( the lower bound varies for different types of groups but is at most 62 ) defined over a finite field of characteristic @xmath4 . it is proved that a finite group @xmath1 isospectral to @xmath3 can not have a nonabelian composition factor which is a group of lie type defined over a field of characteristic distinct from @xmath4 . together with a series of previous results this implies that every finite group @xmath1 isospectral to @xmath3 is ` close ' to @xmath3 . namely , if @xmath3 is a linear or unitary group , then @xmath5 , in particular , there are only finitely many such groups @xmath1 for given @xmath3 . if @xmath3 is a symplectic or orthogonal group , then @xmath1 has a unique nonabelian composition factor @xmath6 and , for given @xmath3 , there are at most 3 variants for @xmath6 ( including @xmath7 ) . groups of lie type , finite classical groups , element orders 20d06 , 20d20
numerical schemes , which is using for solving systems of equations of many - particle dynamics , can have restrictions on a step and an interval of integration because if they increase , the numerical schemes become unstable and do not conserve integrals of motion . as a result , when we simulate many - particle system behaviour on the sufficiently large time interval we should decrease an integration step , which leads to considerable increasing of computation quantity . a motion of @xmath0-particle system in a field with potential @xmath1 can be described by the system of hamiltonian equations @xcite @xmath2 with initial conditions @xmath3 . here @xmath4- particle coordinates , @xmath5- particle momentums , @xmath6 - a dimension of coordinate space and @xmath7 is a separable hamiltonian of the system with symmetric and positive definite mass matrix @xmath8 and a field potential @xmath1 . an approach to design of conservative difference scheme for solving hamiltonian equations ( [ heq ] ) is based on following stages . first is a choice of appropriate type of generating function @xmath9 , which fixes a definite family of simplectic difference scheme @xcite . simplectic difference scheme corresponds to the canonical transformation of canonical variables @xcite . solution of hamiltonian system ( [ heq ] ) for a definite time moment can be represented as canonical transformation and that is why conserves the value of hamiltonian @xcite . second is using of `` forward '' and `` backward '' taylor expansion of coordinate and momentum on a time step for getting of a corresponding generating function and remaining values of coordinate and momentum . at last , obtained at the previous stage explicit and implicit schemes are used for constructing symmetric simplectic scheme for solving hamiltonian equations ( [ heq ] ) . @xmath10 -\ ] ] @xmath11 \cdot\left(\mathbf{q}^{k+1}-\mathbf{q}^k\right),\ ] ] @xmath12-\ ] ] @xmath13 @xmath14 -\ ] ] @xmath15 where @xmath16 is hessian matrix for @xmath17 . obtained by described above approach difference scheme of third order [ 5 ] was tested by solving the system of equations ( [ heq ] ) with hamiltonian @xmath18 so called kepler two body problem , results of numerical calculations are compared for the symmetric simplectic scheme of third order and well - known second order velocity verlet scheme with a time step @xmath19 and a time interval @xmath20 , \,t=5000 $ ] . approximate solution by verlet scheme with a time step h=0.02 is used as an exact solution [ 5 ] . a new approach for constructing symmetric simplectic numerical schemes for solving hamiltonian systems of equations is proposed . the numerical schemes constructed by this approach produce more stable and accurate solution of hamiltonian system ( 1 ) and better conserve the energy of a system on the large interval of numerical integration for a relatively large integration step in comparision with the verlet method , which is usually using for solving equations of motion in molecular dynamics . * acknowledgments . * e.g.n . acknowledge dr . b. batgerel for computations and numerical results . the work was supported in part by the russian foundation for basic research , grant no . 15 - 01 - 06055 .
simulation of many - particle system evolution by molecular dynamics takes to decrease integration step to provide numerical scheme stability on the sufficiently large time interval . it leads to a significant increase of the volume of calculations . an approach for constructing symmetric simplectic numerical schemes with given approximation accuracy in relation to integration step , for solving molecular dynamics hamiltonian equations , is proposed in this paper . numerical experiments show that obtained under this approach symmetric simplectic third order scheme is more stable for integration step , time - reversible and conserves hamiltonian of the system with more accuracy at a large integration interval then second order velocity verlet numerical schemes . + * key words : * hamiltonian systems of equations , simplectic difference schemes , generating functions , molecular dynamics + * pacs : * 02.60.cb 02.60.jh 02.70.bf 02.70.ns
cold atoms have found numerous applications in quantum simulation , quantum chemistry and precision metrology @xcite . for most such experiments it is helpful to trap as many atoms in a magneto - optical trap ( mot ) as possible . in order to achieve this , a collimated slowed atomic beam is often used to load the mot . a common approach to atomic beam slowing makes use of the scattering force from a counter - propagating near - resonant laser beam . in a zeeman slower , the unavoidable doppler shifts are compensated with zeeman shifts from an inhomogeneous applied magnetic field @xcite . this field is commonly produced by a carefully - designed pattern of current - carrying wires around the axis of atomic beam propagation . recently , alternative designs based on permanent magnets have attracted interest @xcite . these use permanent magnets positioned at different distances away from the atomic beam to create the desired field profile . although the resulting field can not be switched on and off , the use of permanent magnets offers numerous practical advantages including robustness , ease of maintenance , the elimination of electrical power and water cooling requirements , and low cost . permanent magnet zeeman slowers can be categorized by the direction of the field relative to the beam axis ( longitudinal or transverse ) and by the mechanism for positioning the individual magnets . for example , in the work of hill _ @xcite , the positions of the magnets are fixed with screws , which can be manually adjusted for fine optimization of the profile of the magnetic field . @xcite realized an automated version of this design by adding servo motors to control and optimize the positions of the magnets . another approach is to position the magnets in a halbach arrangement @xcite . all the above - mentioned designs rely on rigid attachment of magnets to a mechanical construction whose position can be fine - tuned if required . in this paper we present a conceptually novel approach , which relies on `` self - assembly '' of spherical neodymium iron boride ( ndfeb ) magnets into stable structures . the resultant slower is of the longitudinal - field type , meaning that a magnetic field along the atomic beam axis is used to compensate for the doppler shift of the atoms . unlike transverse - field slowers , which use linearly polarized ( @xmath0 ) cooling light , longitudinal slowers use only @xmath1 or only @xmath2 polarized light . as a result we expect our slower to require roughly half the laser power of transverse - field permanent - magnet slowers . the design does not require any supporting mechanical holders , and is stable , quick to assemble , and inexpensive . in section [ design ] we present the general approach and details of the main design considerations . in section [ realization ] we describe the construction of the slower . in section [ results ] we discuss the measured field profile and calculated performance , and we conclude in section [ discussion ] . the basic idea behind our design is to create a longitudinal - field permanent - magnet slower by arranging a large number of individual magnetized spheres in a magnetically stable configuration . design goals include simplicity , low cost , ease of assembly , and tunability ( which can be achieved by adding or subtracting individual magnetic elements ) . previous work has shown that an array of magnetic dipoles can be rigidly positioned so as to create a spatially varying longitudinal field which is suitable for a zeeman slower @xcite . in this paper we concentrate on finding an arrangement of magnetic dipoles which both locally minimizes the magnetic interaction energy among the dipoles and creates a suitable slowing field . we refer informally to such an arrangement as a `` self - assembled zeeman - slower . '' in a zeeman slower , atoms are slowed by an approximately constant deceleration force due to photon scattering from a counterpropagating beam . in the limit of large laser intensity , the maximum deceleration is @xmath3 , where @xmath4 is the wavevector of the slowing light , @xmath5 is the natural linewidth of the relevant optical transition , and @xmath6 is the atomic mass . robust slowing requires operation at some smaller acceleration @xmath7 , where @xmath8 . for constant acceleration , the velocity as a function of position is then @xmath9 where @xmath10 is the initial velocity , @xmath11 is the length of the slower , and @xmath12 is the position along the axis of the slower . to maintain resonance , the doppler shift arising from this variation in velocity must be compensated by the zeeman shift due to an applied magnetic field . for a transition with a magnetic moment of one bohr magneton , the condition for such compensation is @xmath13 , leading to an ideal magnetic field profile of the form @xmath14 figure [ b_ideal ] illustrates a magnetic field profile calculated from equation ( 2 ) with @xmath15 cm and @xmath16 m/s . and @xmath17 . , width=377 ] in our zeeman slower design , a field of the appropriate form is created by a superposition of fields due to individual elements . the components of the field due to a single magnetic dipole placed at the origin and oriented along the @xmath12-axis are @xmath18,\\ & b_{y } = \frac{\mu _ { 0}m}{4\pi } \left [ \frac{3yz}{r^5 } \right],\\ & b_{z } = \frac{\mu _ { 0}m}{4\pi } \left [ \frac{2z^2-x^2-y^2}{r^5 } \right ] , \end{aligned}\ ] ] where @xmath19 is the magnetic moment of the dipole and @xmath20 . the field of our slower is a sum of such terms due to dipoles at different positions . for an array of @xmath21 magnets which has discrete or continuous rotational symmetry about the @xmath12 axis , @xmath22 and @xmath23 are zero at all points on the axis of symmetry , and the total magnetic field on axis is the sum of the @xmath24 components due to each individual magnet : @xmath25 fig . [ lines2 ] illustrates the fields from a combination of 1d arrays of different lengths , with each array starting at the origin . the total field from all three arrays ( solid line ) has the required asymmetric profile . it is easy to see that adding the contributions of more such arrays ( each with incrementally increasing length ) will result in a gradually decreasing total field profile qualitatively resembling that required for a zeeman slower . note that there are two ways to tune the slope of the field profile : by varying the length of each array and also by varying the distance of each array from the @xmath12-axis . the configuration considered in fig . [ lines2 ] has all magnets positioned at the same distance away from the @xmath12-axis . -axis and aligned along it . the spacing between neighboring dipoles is 6 mm , and each dipole moment is 0.075 a@xmath26m@xmath27 . the dashed line corresponds to an array consisting of six such magnets , the dash - dotted to an array of 9 magnets , and the dotted to an array of 12 magnets . the solid line is the sum of the fields of all three arrays.,width=377 ] for simplicity of construction it is important to ensure that the arrangement of dipoles not only generates the correct field profile but also is magnetically stable . although it is possible to fix the position of each dipole with glue or mechanical mounting , here we aim at realizing a stable magnetic configuration which does not require any mechanical support structure . for optimum beam slowing , @xmath22 and @xmath23 should be zero and the longitudinal field should vary as little as possible in the transverse direction , which requires the magnet arrangement to be symmetric around the @xmath12-axis . this means that our structure must be stable in three dimensions . here we briefly consider some basic examples of structures which are stable in one and two dimensions . a stable 1d structure is simply a line of spherical magnets with the south pole of one magnet touching the north pole of the other . in 2d , stability can be achieved for two different patterns ( see fig . [ fig : stable_b_0 ] ) . note that only one of the solutions corresponds to a non - zero total magnetic field at long distances . the 3d case is more complex than 1d and 2d , and supports an infinite number of stable structures . as a practical note , stable 3d arrangements can also be very sensitive to the chronological order in which the dipoles are assembled . one aim of this work was to find a 3d pattern that is stable and results in a zeeman - slower - like profile of the magnetic field . the line of magnets from fig . [ fig : stable_b_0]a can be closed onto itself to make a ring which is very stable and generates no net field through its axis . if one stacks a number of such rings so that their axes overlap the resulting structure will be a cylinder . depending on the relative orientation of magnetization in neighboring rings one can end up with two possible types of regular cylinder . [ fig : cylinder_square_pattern]a illustrates a square lattice which results from each ring being magnetized in the opposite direction to the neighboring ring ( similar to the pattern described in fig . [ fig : stable_b_0]b ) . fig . [ fig : cylinder_square_pattern]b illustrates a triangular lattice , for which each ring is magnetized in the same direction as the neighboring ones ( similar to the pattern described in fig . [ fig : stable_b_0]c ) . both of these cylinders generate no net magnetic field along their axis . although neither of the two cylinders generate any net magnetic field , they play a key role in the design of our zeeman slower . we use the square - patterned cylinder as a rigid bottom layer for a stable arrangement described in the next section . we refer to the inner cylinder as an _ adhesive layer _ , because it keeps the magnets in the outer layers fixed in the correct way . the square - patterned cylinder is especially appropriate for supporting line - shaped outer layers aligned along the length of the cylinder . as discussed in the previous section , we create stable magnetic structures by adding lines of spherical magnets on top of an adhesive cylindrical inner layer . in principle , there is a large number of different arrangements of individual magnets that can generate a zeeman - slower - like field . in choosing the most appropriate structure for a specific application the main constraints arise from the need to keep the whole structure magnetically stable and symmetric around the z - axis , and the need to minimize the total number of individual magnets in order to keep the cost and weight low . for these reasons we chose the shape of the whole structure to have three fin - like elements as shown in fig . [ fig : slower ] . all three fins are identical and magnetized in the same direction and are held in a fixed position by the adhesive layer , which in turn is wound around a non - magnetic tube . the structure and ordering of magnets within each fin is shown in fig . [ fig : fins ] . each individual magnet in the fins is magnetized the same way along the @xmath12-axis . adjacent lines of magnets have incrementally different numbers of elements , and each shorter line is incrementally further away from the z - axis . such a structure results in a slowly decreasing field profile , and three such fins produce a @xmath12-directed axial field three times larger than than that from any one fin . as seen in fig . [ fig : slower ] , all fins are positioned symmetrically around the z - axis at 120 degrees to each other . each individual magnet used in this assembly is a 6 mm diameter n45 grade spherical neodymium iron boride magnet , with a dipole moment of about 0.075 a@xmath26m@xmath27 . [ fig : slowerfield ] presents the measured longitudinal magnetic field along the @xmath12-axis of the slower . the slowing region is from around 15 to 30 cm , resulting in a 15 cm slowing length . this is sufficient for zeeman slowing of strontium , due to the broad ( @xmath2832 mhz ) linewidth and consequent high scattering rate of this atom s 461 nm transition . this design could also be straightforwardly adapted to slowing of other atomic species . the field is reasonably smooth and can be tweaked and improved further by moving the individual magnets around the outer layers of the fins . variation of the axial field as the position is scanned in @xmath29 and @xmath30 is plotted in the inset of fig . [ fig : slowerfield ] . these data were taken at the point of maximum field , near the slower entrance . the variation of the zeeman shift across a typical atomic beam diameter is small compared to the linewidth of the relevant transition in strontium , indicating that transverse variations in the axial field should not significantly limit slower performance . . inset plots measured variations in the axial field at the maximum field point as the position is scanned transverse to the axis . , width=453 ] using the measured field profile one can integrate the equations of motion to determine the fate of atoms with different incoming velocities . atoms traveling at velocity @xmath31 through a magnetic field @xmath32 experience both a doppler shift @xmath33 and a zeeman shift @xmath34 , where @xmath35 is the wavelength of the slowing laser beam and @xmath36 is the magnetic moment of the relevant atomic transition . together with the chosen detuning @xmath37 of the laser from the atomic transition , the sum of these terms sets the local effective detuning at any point along the @xmath12 axis : @xmath38 the average spontaneous light pressure force is then given by @xmath39 where @xmath40 is the local saturation parameter , which we choose to be constant for simplicity . the resulting equation of motion can be integrated to find the velocity as a function of time or distance for different starting velocities . [ fate ] presents atomic trajectories for different initial velocities in @xmath41 parameter space for the measured field profile of our zeeman slower . the highest captured velocity is limited to around @xmath42 m / s , which could be increased by simply adding magnets to increase the maximum magnetic field . note that the atoms are significantly decelerated in the region between 0.3 and 0.4 meters despite the opposite slope of the magnetic field . this can be explained by the large natural linewidth ( @xmath43 mhz ) of strontium s 461 nm line , which leads to substantial photon scattering rates even at relatively large effective detunings . taking this into account , by varying the saturation parameter and laser detuning it is possible to produce a final velocity of around 25 m / s , which is below the capture velocity of a typical mot . for similar reasons , the atoms with initial velocities below @xmath44 m / s are decelerated too early and reach zero velocity before reaching the end of the slower . these atoms are therefore lost since they will never reach the mot region and can not be trapped . however , for a typical oven temperature of 600@xmath45c , the most probable longitudinal velocity of the atomic beam is close to 500 m / s . this means that there are significantly more atoms with initial velocities between 170 and 370 m / s than with velocities between 0 and 170 m / s , even before taking into account scattering processes which can preferentially deplete the population of low - velocity atoms @xcite . therefore this slower design is capable of trapping more than 75% of the atoms with starting velocities below 370 m / s . of 1.5 and a detuning @xmath37 of 340 mhz.,width=453 ] we have presented a new type of longitudinal zeeman slower based on spherical permanent magnets . in particular we demonstrated how one can assemble such magnets into a stable structure which provides an axial magnetic field profile suitable for zeeman slowing . like other permanent magnet slowers , this design does not require electrical current or water cooling . the proposed design is flexible , easy to assemble , and inexpensive . the authors thank ruwan senaratne , shankari rajagopal , and zach geiger for useful discussions and physical insights . we acknowledge support from the air force office of scientific research ( award fa9550 - 12 - 1 - 0305 ) and the alfred p. sloan foundation ( grant br2013 - 110 ) . 99 w. d. phillips , `` laser cooling and trapping of neutral atoms '' , reviews of modern physics * 70 , * 721741 ( 1998 ) c. e. wieman , d. e. pritchard and d. j. wineland , `` atom cooling , trapping , and quantum manipulation '' , reviews of modern physics * 71 , * s253s262 ( 1999 ) i. bloch , j. dalibard and s. nascimbene , `` quantum simulations with ultracold quantum gases '' , nature physics * 8 , * 267 - 276 ( 2012 ) d. s. jin and j. ye , `` introduction to ultracold molecules : new frontiers in quantum and chemical physics '' , chemical reviews * 112 , * 48014802 ( 2012 ) a. derevianko and h. katori , `` _ colloquium _ : physics of optical lattice clocks '' , reviews of modern physics * 83 , * 331347 ( 1998 ) b. j. bloom et al , `` an optical lattice clock with accuracy and stability at the @xmath46 level '' , nature * 506 * , 7175 ( 2014 ) w. d. phillips and h. metcalf , `` laser deceleration of an atomic beam '' , physical review letters * 48 , * 596599 ( 1982 ) y. b. ovchinnikov , `` a zeeman slower based on magnetic dipoles '' , optics communications * 276 , * 261267 ( 2007 ) y. b. ovchinnikov , `` longitudinal zeeman slowers based on permanent magnetic dipoles '' , optics communications * 285 , * 11751180 ( 2012 ) i. r. hill et al , `` a simple , configurable , permanent magnet zeeman slower for strontium '' , eftf 2012 - 2012 european frequency and time forum , proceedings , art . 6502439 , pp . 545 - 549 ( 2012 ) p. cheiney et al , `` a zeeman slower design with permanent magnets in a halbach configuration '' , reviews of scientific instruments * 82 , * 063115 ( 2011 ) g. reinaudi , c. b. osborn , k. bega , and t. zelevinsky , `` dynamically configurable and optimizable zeeman slower using permanent magnets and servomotors '' , j. opt . b * 29 , * 729 ( 2012 ) i. r. hill et al , `` zeeman slowers for strontium based on permanent magnets '' , j. phys . b : at . mol . . phys . * 47 * , 075006 ( 2014 )
we present a novel type of longitudinal zeeman slower . the magnetic field profile is generated by a 3d array of permanent spherical magnets , which are self - assembled into a stable structure . the simplicity and stability of the design make it quick to assemble and inexpensive . in addition , as with other permanent magnet slowers , no electrical current or water cooling is required . we describe the theory , assembly , and testing of this new design .
it is well - known that observability and detectability are fundamental concepts in system analysis and synthesis ; see , e.g. , @xcite . in the linear system theory , detectability is a weaker concept than observability , since it describes the fact that all unobservable states are asymptotically stable . over the last two decades , the classical detectability in the linear system theory has been extended to stochastic systems in different ways . for example , the definition of stochastic detectability for time - invariant it stochastic systems can be found in @xcite , which is dual to mean square stabilization . in @xcite , the notions of exact observability and exact detectability were presented for it stochastic systems , which led to the stochastic popov - belevitch - hautus ( pbh ) criteria like those for deterministic systems . another natural concept of detectability for it stochastic systems was given in @xcite based on the idea that any non - observed states corresponded to stable models of the system . in @xcite , the exact detectability in @xcite and detectability in @xcite were proved to be equivalent , and a unified treatment was proposed for detectability and observability of it stochastic systems . based on the standard notions of detectability and observability for time - varying linear systems @xcite , studied in @xcite were detectability and observability of discrete time - invariant stochastic systems as well as the properties of lyapunov equations . recently , the exact detectability and observability were extended to stochastic systems with markov jumps and multiplicative noise in @xcite . as it is well known that the classical lyapunov theorem is very essential in stability theory , which asserts that if a matrix @xmath6 is schur stable , then for any @xmath7 , the classical lyapunov equation @xmath8 admits a unique solution @xmath9 ; conversely , if @xmath10 is detectable , @xmath7 , and the lyapunov equation @xmath8 admits a unique solution @xmath9 , then @xmath6 is schur stable . the classical lyapunov theorem was generalized to deterministic time - varying systems in @xcite and will be extended to stochastic time - varying systems in this paper under any one assumption of uniform detectability , @xmath11-exact detectability and @xmath11-exact observability . recently it has become known that discrete - time stochastic systems with multiplicative noise are ideal models in the fields of investment portfolio optimization @xcite , system biology @xcite and so on . so the discrete - time stochastic @xmath12 control and filtering design have been extensively studied in recent years ; see , e.g. , @xcite and the references therein . as it is well - known , time - varying systems may be utilized to model more realistic systems and are more challenging in mathematics than time - invariant ones . so far , the majority of the existing results is focused on detectability of time - invariant systems only , except for a few about time - varying systems ; see @xcite . because linear time - invariant systems are not sufficient to describe many practical phenomena , this motivates researchers to study time - varying systems . in the classical work @xcite , uniform detectability of the deterministic linear discrete - time time - varying ( ldtv ) system @xmath13 was defined and discussed . by the duality of stochastic stabilizability , another definition called stochastic detectability " was introduced in @xcite for ldtv markov systems , which is not equivalent to uniform detectability in time - varying case . mainly motivated by the preceding discussion and the authors series works @xcite , this paper will study detectability , observability and lyapunov - type equation related to ldtv stochastic systems with multiplicative noise . firstly , the classical uniform detectability of @xcite for such systems is extended , and some properties on uniform detectability are obtained . by means of our lemma [ lem:2.1 ] given later , we obtain the observability gramian matrix @xmath14 and the state transition matrix @xmath15 , which are deterministic matrices and easy to be applied in practice . specifically , we prove an important theorem that uniform detectability preserves invariance under an output feedback control law , which is expected to be useful in stochastic @xmath12 control . as an application , under the assumption of uniform detectability , lyapunov - type theorems on stochastic stability are also presented . secondly , we extend exact detectability of linear continuous - time stochastic it systems @xcite to ldtv systems . we introduce four concepts called @xmath3-exact detectability , @xmath2-exact detectability , @xmath1-exact detectability and @xmath16-exact detectability , and they in turn become weaker in the sense that the former implies the latter in a sequence . although in linear time - invariant system @xmath17 these four concepts are equivalent with @xmath18 , but they are different from the others in the time - varying case , which reveals the essential difference between time - invariant and time - varying systems . it is shown that uniform detectability implies @xmath19-exact detectability ( see lemma [ lem:3.1.3 ] ) , and stochastic detectability @xcite implies the above four types of exact detectability ( see proposition [ prop:3.1.1 ] and remark [ rem:3.1.3 ] ) . it seems that there is no inclusion relation among uniform detectability , @xmath3-exact detectability , @xmath2-exact detectability and @xmath20-exact detectability , although they can be unified in the linear discrete time - invariant systems @xcite . two important lyapunov - type theorems under @xmath3-exact detectability for periodic systems are obtained ( see theorems [ thm:5.3.1][thm:5.3.2 ] ) , which reveal the important relation between the exponential stability and the existence of positive definite solutions of generalized lyapunov equations ( gles ) . parallel to various definitions on exact detectability , we also introduce @xmath3-exact observability , @xmath2-exact observability , @xmath1-exact observability and @xmath19-exact observability , which are respectively stronger than @xmath3-exact detectability , @xmath2-exact detectability , @xmath1-exact detectability and @xmath19-exact detectability . for the linear time - invariant system ( [ eq vgthgh ] ) , @xmath21- , @xmath2- , @xmath20- and @xmath16-exact observability are equivalent , but they are different definitions for the linear time - varying system ( [ eq vfhv ] ) . we present a rank criterion for @xmath19- and a criterion for @xmath3-exact observability based on the gramian matrix @xmath22 . finally , under the assumption of @xmath3-exact observability , a lyapunov - type theorem is derived from theorem [ thm:5.3.1 ] . the rest of the paper is organized as follows . in section [ sec : ud ] , we define uniform detectability and discuss its properties . lyapunov - type theorems are given under uniform detectability . section [ sec : edeo ] introduces some new concepts about exact detectability and exposes nice properties . this section also presents lyapunov - type stability theorems based on @xmath5-exact detectability . moreover , the relation among uniform detectability , exact detectability and stochastic detectability is clarified via some examples . section [ sec : tltt ] introduces various definitions for exact observability , which are stronger than those of section-[sec : edeo-1 ] . section [ sec : ltted ] provides some comments on this study . finally , section [ sec : concl ] concludes the paper with some remarks . _ notation : _ @xmath23 : the set of all real @xmath24-dimensional vectors . @xmath25 : the set of all @xmath26 symmetric matrices whose entries may be complex numbers . @xmath27 : the set of all complex numbers . @xmath28 : the set of all @xmath29 real matrices . @xmath30 : the norm of a vector or matrix . @xmath31 ( resp . @xmath32 ) : @xmath33 is a real symmetric positive definite ( resp . positive semi - definite ) matrix . @xmath34 : the identity matrix . @xmath35 : the spectrum set of the operator or matrix @xmath36 . @xmath37 : the transpose of matrix @xmath33 . @xmath38 , especially , @xmath39 , @xmath40 . @xmath41 . in this section , we will define one important concept for ldtv stochastic systems , called uniform detectability " . and then , we will obtain lyapunov - type theorems under uniform detectability , which are extensions of classical lyapunov theorem . consider the following ldtv stochastic system @xmath42 where @xmath43 is the @xmath24-dimensional state vector , @xmath44 is the @xmath45-dimensional measurement output , @xmath46 represents a one - dimensional independent white noise process defined on the filtered probability space @xmath47 with @xmath48 . assume that @xmath49 , @xmath50=\delta_{kj}$ ] , where @xmath51 is a kronecker function defined by @xmath52 for @xmath53 while @xmath54 for @xmath55 . @xmath56 is assumed to be deterministic for simplicity purposes , and @xmath57 , @xmath58 and @xmath59 are time - varying matrices of appropriate dimension . in practice , one is more concerned about the @xmath60-solution @xmath61 of stochastic difference equation @xmath62 [ def:2.1]the stochastic vector - valued sequence @xmath63 is called a solution of system ( [ eq vhvhc ] ) if ( i ) @xmath64 ; ( ii ) @xmath65 solves ( [ eq vhvhc ] ) for @xmath66 ; ( iii ) @xmath67 , where @xmath68 is assumed to be a trivial sigma algebra . system ( [ eq vhvhc ] ) is said to have a unique solution if for any two of its solutions @xmath69 and @xmath70 , @xmath71 . [ rem:2.1]it can be found that , in most present literature , the condition ( iii ) in definition [ def:2.1 ] is not particularly pointed out when defining solutions of system ( [ eq vhvhc ] ) , which is in fact an essential requirement as done in stochastic differential equations @xcite . this makes an fundamental difference of ( [ eq vhvhc ] ) from deterministic difference equations , as will be seen in the following examples . [ exm:2.1]it is easy to see that the following forward difference equation @xmath72 always admits a unique solution on @xmath73 $ ] . in addition , if @xmath74 are nonsingular , then the backward difference equation @xmath75 also has a unique solution on @xmath73 $ ] . [ exm:2.2]obviously , the linear stochastic difference equation ( [ eq vhvhc ] ) always has a unique @xmath76-solution @xmath43 on any interval @xmath77 $ ] . however , even if @xmath74 , are nonsingular , the following stochastic difference equation @xmath78 with terminal state given does not always admit an @xmath79-solution . for example , if we take @xmath80 , @xmath81 , and the terminal state @xmath82 in ( [ eq fdfdf ] ) , then @xmath83 , @xmath84 . [ rem:2.2]a class of backward stochastic difference equations arising from the study of discrete stochastic maximum principle can be found in @xcite . to define and better understand the uniform detectability for system ( [ eq vcgtgh1 ] ) , we first give some lemmas . [ lem:2.2](i ) for system ( [ eq vcgtgh1 ] ) , @xmath85 for @xmath86 , where it is assumed that @xmath87 , and @xmath15 is given by the following iterative relation @xmath88 , \ l > k.\ ] ] ( ii ) @xmath89 if @xmath90 and @xmath91 are bounded for @xmath92 . \(i ) can be shown by induction . for @xmath93 , we have @xmath94\nonumber \\ & = e[x_{l-1}^t(f^t_{l-1}f_{l-1}+g_{l-1}^tg_{l-1})x_{l-1}]\nonumber\\ & = e\|\phi_{l , l-1}x_{l-1}\|^2 . \nonumber\end{aligned}\ ] ] hence , ( [ eq 1 ] ) holds for @xmath93 . assume that for @xmath95 , @xmath96 . next , we prove @xmath97 it can be seen that @xmath98\nonumber \\ & = e[(f_{m-1}x_{m-1}+g_{m-1}x_{m-1}w_{m-1})^t\phi_{l , m}^t\phi_{l , m}(f_{m-1}x_{m-1}+g_{m-1}x_{m-1}w_{m-1})]\nonumber\\ & = e[x_{m-1}^t(f_{m-1}^t\phi_{l , m}^t\phi_{l , m}f_{m-1}+g_{m-1}^t\phi_{l , m}^t\phi_{l , m}g_{m-1})x_{m-1}]\nonumber\\ & = e\|\phi_{l , m-1}x_{m-1}\|^2.\label{eq step1 in lemma 1}\nonumber\end{aligned}\ ] ] this completes the proof of ( i ) . and ( ii ) is obvious . the proof of this lemma is complete . [ lem:2.1]for system ( [ eq vcgtgh1 ] ) , there holds @xmath99 for @xmath100 , where @xmath101 \label{eq 2}\ ] ] with @xmath102 and @xmath103 given by ( [ eq 1 ] ) . based on lemmas [ lem:2.2][lem:2.1 ] , we are now in a position to define the uniform detectability for system ( [ eq vcgtgh1 ] ) . [ def:2.2]system ( [ eq vcgtgh1 ] ) or @xmath111 is said to be uniformly detectable if there exist integers @xmath112 , and positive constants @xmath113 , @xmath114 with @xmath115 and @xmath116 such that whenever @xmath117 there holds @xmath118 where @xmath119 , and @xmath120 and @xmath121 are the same as defined in lemma [ lem:2.1 ] . obviously , without loss of generality , in definition [ def:2.2 ] we can assume that @xmath122 . by lemmas [ lem:2.2][lem:2.1 ] , the uniform detectability of @xmath111 implies , roughly speaking , that the state trajectory decays faster than the output energy does . in what follows , @xmath123 is called an observability gramian matrix , and @xmath15 a state transition matrix from @xmath43 to @xmath124 of stochastic system ( [ eq vcgtgh1 ] ) . so ( [ eq vhhjj1 ] ) can be written as @xmath125\ge b^2 e\|x_k\|^2 $ ] . if @xmath126 for @xmath127 , then system ( [ eq vcgtgh1 ] ) reduces to the following deterministic system @xmath128 which was discussed in @xcite . similarly , uniform observability can be defined as follows : [ def : uo]system ( [ eq vcgtgh1 ] ) or @xmath111 is said to be uniformly observable if there exist an integer @xmath129 and a positive constant @xmath130 such that @xmath131 holds for each initial condition @xmath132 , @xmath119 . [ rem : uo]different from the uniform detectability concept , uniform observability needs that any models ( unstable and stable ) should be reflected by the output . this section concentrates on the uniform detectability of system ( [ eq vcgtgh1 ] ) , since it is weaker than uniform observability . uniform observability is also an important concept , which will be further studied in the future . [ def:2.3]system ( [ eq vcgtgh1 ] ) is said to be exponentially stable in mean square ( esms ) if there exist @xmath133 and @xmath134 such that for any @xmath135 , there holds @xmath136 [ prop:2.1]if system ( [ eq vcgtgh1 ] ) is esms , then for any bounded matrix sequence @xmath137 , system ( [ eq vcgtgh1 ] ) is uniformly detectable . by definition [ def:2.3 ] , for any @xmath138 , we always have @xmath139 by ( [ eq vcgvff ] ) , @xmath140 as @xmath141 . set a large @xmath142 such that @xmath143 . then , for any fixed @xmath144 , ( [ eq vhbjh1 ] ) holds only for @xmath145 , which makes ( [ eq vhhjj1 ] ) valid for any @xmath146 and @xmath130 with an equality . so system ( [ eq vcgtgh1 ] ) is uniformly detectable . [ rem:2.3]for system ( [ eq vdchcf ] ) , definition [ def:2.2 ] reduces to definition 2.1 in @xcite . it is easy to prove that uniform detectability coincides with classical detectability of the linear time - invariant system ( [ eq vgthgh ] ) . the following lemma will be used throughout this paper . [ lem:2.3]for a nonnegative real sequence @xmath147 , if there exist constants @xmath148 , @xmath149 , and an integer @xmath150 such that @xmath151 and @xmath152 , then @xmath153(\delta_0^{h_0})^{k - k_0}s_{k_0 } , \ \forall k\ge k_0.\ ] ] the following proposition extends lemma 2.2 in @xcite . [ prop:2.2]suppose that @xmath111 is uniformly detectable , and @xmath57 and @xmath58 are uniformly bounded , i.e. , @xmath154 . then @xmath155 implies @xmath156 . if there exists some integer @xmath157 such that for all @xmath158 , @xmath159 , then @xmath160 @xmath161 . moreover , @xmath162\le 2m^2e\|x_i\|^2\le m_0e\|x_i\|^2 $ ] , where @xmath163 . by lemma [ lem:2.3 ] , not only does @xmath156 , but also is system ( [ eq vcgtgh1 ] ) esms . otherwise , there exists a subsequence @xmath164 such that @xmath165 . now , for @xmath166 , we write @xmath167 with @xmath168 , then @xmath169 @xmath170 @xmath171 therefore , we have @xmath172 obviously , in order to show @xmath173 , we only need to show @xmath174 . if it is not so , then there are a subsequence @xmath175 of @xmath176 and @xmath177 , such that @xmath178 , @xmath179 . by definition [ def:2.2 ] , @xmath180\ge b^2e\|x_{n_i}\|^2>b^2\varsigma . \label{eq vhhbb}\ ] ] taking @xmath181 in ( [ eq vhhbb ] ) , we have @xmath182 , which is a contradiction . hence , the proof is complete . in the remainder of this section , we will prove the output feedback invariance for uniform detectability . consider the following ldtv stochastic control system @xmath183 applying an output feedback control law @xmath184 to ( [ eq vcgtgh1dd ] ) yields the following closed - loop system @xmath185 [ thm:2.1]if @xmath111 is uniformly detectable , then so is @xmath186 . by lemma [ lem:2.1 ] , the observability gramian for system ( [ eq vcgfdef ] ) is @xmath187 , where @xmath188 , \ \bar \phi_{k+i , k}=\left [ \begin{array}{cc } \bar \phi_{k+i , k+1}\bar f_k\\ \bar \phi_{k+i , k+1}\bar g_k \end{array } \right ] , \ i=1,\cdots , s.\ ] ] @xmath189 to prove that @xmath190 is uniformly detectable , it suffices to show that there are constants @xmath191 , @xmath192 , @xmath112 such that for @xmath193 , @xmath119 , whenever @xmath194<{\bar b}^2 e\|x_k\|^2 , \label{eq vhhjj}\ ] ] we have @xmath195 it is easy to show @xmath196.\ ] ] where * represents terms involving @xmath197 , @xmath198 , @xmath199 and @xmath200 , @xmath201 . hence , for any @xmath132 , @xmath202 \le e[x_k^t\bar { \cal o}_{k+s , k}x_k ] \le \varrho e[x_k^t { \cal o}_{k+s , k}x_k ] , \label{eq cvghv}\end{aligned}\ ] ] where @xmath203 , @xmath204 . in addition , by observation , for any @xmath205 , @xmath206 where @xmath207 is a matrix involving @xmath197 , @xmath198 , @xmath199 and @xmath200 , @xmath208 . if we take @xmath209 , then it follows from ( [ eq cvghv ] ) that @xmath210<\frac 1 \rho e[x_k^t\bar { \cal o}_{k+s , k}x_k]\le \frac { \bar b^2}{\rho } e\|x_k\|^2\le b^2e\|x_k\|^2 $ ] . by the uniform observability of @xmath111 , it follows that @xmath211\nonumber\\ & \le \big(2d^2 + 2\mu^2\frac { \bar b^2 } { \rho}\big)e\|x_k\|^2\nonumber\\ & = \bar d e\|x_k\|^2 , \end{aligned}\ ] ] where @xmath212 , @xmath213 . if we take @xmath214 to be sufficiently small , then @xmath215 , which yields the uniform detectability of @xmath216 . hence , the proof of this theorem is complete . theorem [ thm:2.1 ] reveals that the output feedback does not change uniform detectability . [ exm:2.4]for simplicity , we set @xmath217 . then it can be computed that @xmath218=\left [ \begin{array}{cc } h_k \\ { h_{k+1}}(f_k+m_kk_kh_k)\\ { h_{k+1}}(g_k+n_kk_kh_k ) \end{array } \right],\ ] ] @xmath219=\left [ \begin{array}{cc } h_k \\ { h_{k+1}}g_k \end{array } \right].\ ] ] obviously , @xmath220.\ ] ] [ exm:2.5]by definition , we have @xmath221 , \ \ { \phi}_{k+1,k}=\left [ \begin{array}{ccc } f_k\\ g_k \end{array } \right].\ ] ] hence , @xmath222 with @xmath223 . $ ] in the following , we will further study the following time - varying gle @xmath224 under uniform detectability . the aim of this subsection is to extend the classical lyapunov theorem to gle ( [ eq vhvcxf ] ) . to study ( [ eq vhvcxf ] ) , we first introduce the following finite time backward difference equation @xmath225 obviously , equation ( [ eq vceqwe ] ) has nonnegative definite solutions @xmath226 . [ prop:4.1.1]@xmath227 is monotonically increasing with respect to @xmath228 , i.e. , for any @xmath229 , @xmath230 [ prop:4.1.2]if system ( [ eq vcgtgh1 ] ) is esms , and @xmath59 is uniformly bounded ( i.e. , there exists @xmath242 such that @xmath243 , @xmath244 ) , then the solution @xmath227 of ( [ eq vceqwe ] ) is uniformly bounded for any @xmath245 and @xmath246 $ ] . by ( [ eq hvbhcdgt ] ) , for any deterministic @xmath247 , we have @xmath248\le \sum_{i = k}^{\infty } e[x_i^th_i^th_ix_i]\\ & \le m^2\|x_k\|^2\beta \sum_{i = k}^{\infty}\lambda^{(i - k)}=m^2\|x_k\|^2 \beta \frac{1}{1-\lambda } , \end{aligned}\ ] ] which leads to that @xmath249 since @xmath250 is arbitrary . hence , the proof is complete . combining proposition [ prop:4.1.1 ] with proposition [ prop:4.1.2 ] yields that @xmath251 exists , which is a solution of ( [ eq vhvcxf ] ) . hence , we obtain the following lyapunov - type theorem . [ thm:4.1.1]if system ( [ eq vcgtgh1 ] ) is esms and @xmath252 is uniformly bounded , then ( [ eq vhvcxf ] ) admits a unique nonnegative definite solution @xmath253 . the converse of theorem [ thm:4.1.1 ] still holds . [ thm:4.1.2]suppose that @xmath111 is uniformly detectable and @xmath57 and @xmath58 are uniformly bounded with an upper bound @xmath242 . if there is a bounded nonnegative definite symmetric matrix sequence @xmath254 solving gle ( [ eq vhvcxf ] ) , then system ( [ eq vcgtgh1 ] ) is esms . for system ( [ eq vcgtgh1 ] ) , we take a lyapunov function as @xmath255 where @xmath256 is to be determined . for simplicity , in the sequel , we let @xmath257 . it is easy to compute @xmath258-e[x_{k+1}^t(p_{k+1}+\varepsilon i)x_{k+1}]\nonumber\\ & = e[x_k^t(p_k+\varepsilon i)x_k]-e[(f_kx_k+g_kx_kw_k)^t(p_{k+1 } + \varepsilon i ) ( f_kx_k+g_kx_kw_k)]\nonumber\\ & = e[x_k^t(p_k - f^t_kp_{k+1}f_k - g_k^tp_{k+1}g_k)x_k]+\varepsilon e[x_k^t(i - f_k^tf_k - g_k^tg_k)x_k]\nonumber\\ & = e\|y_k\|^2+\varepsilon e[x_k^t(i - f_k^tf_k - g_k^tg_k)x_k]\nonumber\\ & = e\|y_k\|^2+\varepsilon e\|x_k\|^2-\varepsilon e\|x_{k+1}\|^2 . \label{eq vhbh}\end{aligned}\ ] ] identity ( [ eq vhbh ] ) yields @xmath259+[ev_{k+1}-ev_{k+2}]+\cdots+[ev_{k+s}-ev_{k+s+1}]\nonumber\\ & = \sum_{i = k}^{k+s}e\|y_i\|^2+\varepsilon e\|x_k\|^2-\varepsilon e\|x_{k+s+1}\|^2 . \label{eq fsdfd}\end{aligned}\ ] ] when @xmath260 , we first note that @xmath261 then , by ( [ eq fsdfd ] ) , we still have @xmath262e\|x_k\|^2 . \label{eq vhvhn}\end{aligned}\ ] ] from ( [ eq vhvhn ] ) , it readily follows that @xmath263\}e\|x_k\|^2\nonumber\\ & \le \left\{1-\frac{[b^2+\varepsilon[1-(2m^2)^{s+1}]]}{\lambda_{max}(p_k+\varepsilon i)}\right\}ev_k . \label{eq cvgvg}\end{aligned}\ ] ] considering that @xmath264 is uniformly bounded , if @xmath265 is taken to be sufficiently small , then there must exist a @xmath266 such that @xmath267 when @xmath268 , by uniform detectability we have @xmath269 . from ( [ eq fsdfd ] ) , it follows that @xmath270 similarly , we can show that there exists a constant @xmath271 such that @xmath272 set @xmath273 , in view of ( [ eq vvhbb ] ) and ( [ eq vcdcdv ] ) , we have @xmath274 from identity ( [ eq vhbh ] ) , we know @xmath275 taking @xmath276 in ( [ eq vhvhhk ] ) , it is easy to derive that there exists a positive constant @xmath148 satisfying @xmath277 applying lemma [ lem:2.3 ] with @xmath278 , @xmath279 , @xmath280 $ ] , @xmath281 , it follows that @xmath282 which implies that system ( [ eq vcgtgh1 ] ) is esms due to the fact that @xmath254 is uniformly bounded . the above theorem directly yields the following result . [ coro:4.1.3]suppose that there exists @xmath283 such that @xmath284 for @xmath285 . additionally , if there is a uniformly bounded symmetric matrix sequence @xmath286 solving gle ( [ eq vhvcxf ] ) , then system ( [ eq vcgtgh1 ] ) is esms . we recall that for the linear time - invariant system @xmath287 its exact observability was defined in @xcite , while the same definition for linear continuous - time time - invariant it systems was given in @xcite . for the ldtv stochastic system ( [ eq vcgtgh1 ] ) , the complete observability that is different from the uniform observability @xcite was defined in @xcite . in this section , we will study exact detectability of the stochastic system ( [ eq vcgtgh1 ] ) , from which it can be found that there are some essential differences between the time - varying and time - invariant systems . in addition , lyapunov - type theorems are also presented . we first give several definitions . [ def:3.1.1]for system ( [ eq vcgtgh1 ] ) , @xmath288 is called a @xmath289-unobservable state if @xmath290 a.s . for @xmath291 , and @xmath288 is called a @xmath292-unobservable state if @xmath290 a.s . for @xmath293 $ ] . [ rem:3.1.1]from definition [ def:3.1.1 ] , we point out the following obvious facts : ( i ) if @xmath235 is a @xmath289-unobservable state , then for any @xmath294 , it must be a @xmath292-unobservable state ; ( ii ) if @xmath235 is a @xmath295-unobservable state , then for any @xmath296 , it must be a @xmath292-unobservable state . [ exm:3.1.1]in system ( [ eq vcgtgh1 ] ) , if we take @xmath297 for @xmath298 , then any state @xmath288 is a @xmath289-unobservable state . for any @xmath299 , @xmath300 is a trivial @xmath289-unobservable state . different from the linear time - invariant system ( [ eq vcgdsdsf ] ) , even if @xmath301 is a @xmath302-unobservable state , @xmath303 may not be a @xmath304-unobservable state for any @xmath305 , which is seen from the next example . [ exm:3.1.2]consider the deterministic linear time - varying system with @xmath81 and @xmath306 , \ \mbox{if } \ k \ \mbox{is } \ \mbox{even},\\ \left [ \begin{array}{ccc } 0 & 0\\ 0 & 1 \end{array } \right ] , \ \mbox{if } \ k \ \mbox{is } \ \mbox{odd}. \end{array } \right.\ ] ] obviously , @xmath307 $ ] is a @xmath308-unobservable state due to @xmath309 for @xmath127 , but @xmath310 $ ] is not a @xmath311-unobservable state for any @xmath305 due to @xmath312\ne 0 $ ] , let alone @xmath313-unobservable state . [ def:3.1.2]system ( [ eq vcgtgh1 ] ) is called @xmath289-exactly detectable if all @xmath289-unobservable state @xmath314 is exponentially stable , i.e. , there are constants @xmath133 , @xmath315 such that @xmath316 similarly , system ( [ eq vcgtgh1 ] ) is called @xmath292-exactly detectable if ( [ eq vcgvass ] ) holds for all @xmath292-unobservable state @xmath314 . [ def:3.1.3]system ( [ eq vcgtgh1 ] ) ( or @xmath111 ) is said to be @xmath16-exactly detectable if it is @xmath317-exactly detectable for any @xmath127 . if there exists a nonnegative integer sequence @xmath318 with the upper limit @xmath319 such that system ( [ eq vcgtgh1 ] ) is @xmath320-exactly detectable , i.e. , for any @xmath320-unobservable state @xmath321 , @xmath322 then system ( [ eq vcgtgh1 ] ) ( or @xmath111 ) is said to be weakly finite time or @xmath1-exactly detectable . if @xmath323 , then system ( [ eq vcgtgh1 ] ) ( or @xmath111 ) is said to be finite time or @xmath2-exactly detectable . a special case of @xmath2-exact detectability is the so - called @xmath3-exact detectability , which will be used to study gles . [ def:3.1.4]if there exists an integer @xmath324 such that for any time @xmath325 , system ( [ eq vcgtgh1 ] ) ( or @xmath111 ) is @xmath326-exactly detectable , then system ( [ eq vcgtgh1 ] ) ( or @xmath111 ) is said to be @xmath3-exactly detectable . from definitions [ def:3.1.3][def:3.1.4 ] , we have the following inclusion relation @xmath327 in this paper , we will mainly use @xmath0- and @xmath3-exact detectability . obviously , @xmath3-exact detectability implies @xmath0-exact detectability , but the converse is not true . we present the following examples to illustrate various relations among several definitions on detectability . for illustration simplicity , we only take the concerned examples to be deterministic . [ exm:3.1.3]in system ( [ eq vdchcf ] ) , we take @xmath80 for @xmath127 , and @xmath328 in this case , system ( [ eq vdchcf ] ) ( or @xmath329 ) is @xmath4-exactly detectable , and the zero vector is the unique @xmath330-unobservable state . @xmath329 is also @xmath20-exactly detectable , where @xmath331 . however , @xmath329 is not @xmath2-exactly detectable , and , accordingly , is not @xmath3-exactly detectable for any @xmath332 . [ exm:3.1.4]in system ( [ eq vdchcf ] ) , if we take @xmath80 and @xmath333 for @xmath127 , then @xmath329 is @xmath11-exactly detectable for any @xmath324 , but @xmath329 is not uniformly detectable . this is because for any @xmath334 , @xmath335 and @xmath336 , we always have @xmath337 . but there do not exist @xmath130 and @xmath129 satisfying ( [ eq vhhjj1 ] ) , because @xmath338 while @xmath339 . [ exm:3.1.5]in system ( [ eq vdchcf ] ) , if we take @xmath80 for @xmath127 , and @xmath340 and @xmath341 for @xmath342 , then @xmath329 is uniformly detectable and @xmath343-exactly detectable , but it is not @xmath344-exactly detectable . the following lemma is obvious . [ lem:3.1.1]at any time @xmath157 , @xmath300 is not only a @xmath289- but also a @xmath292-unobservable state for any @xmath294 . by lemma [ lem:3.1.1 ] , if we let @xmath345 denote the set of all the @xmath302-unobservable states of system ( [ eq vcgtgh1 ] ) at time @xmath157 , then @xmath345 is not empty . furthermore , it is easy to show that @xmath345 is a linear vector space . [ lem:3.1.2]for @xmath346 , if there does not exist a nonzero @xmath347 such that @xmath348 , @xmath349 , @xmath350 , then @xmath290 a.s . with @xmath158 implies @xmath300 a.s .. from @xmath351 a.s . , it follows that @xmath352=0 . \label{eq gfff}\ ] ] from @xmath353 a.s . , @xmath354 , @xmath355 , it follows from lemma [ lem:2.1 ] that @xmath356=0 . \label{eq vghvh}\ ] ] let @xmath357 $ ] , rank@xmath358 . when @xmath359 , this implies @xmath300 a.s . , and this lemma is shown . for @xmath360 , by the result of @xcite , there are real nonzero vectors @xmath361 such that @xmath362 . by ( [ eq gfff ] ) , we have @xmath363 & = \mathrm{trace}e[h^t_{k_0}h_{k_0}x_{k_0}x_{k_0}^t]\nonumber\\ & = \mathrm{trace}\{h^t_{k_0}h_{k_0}e[x_{k_0}x_{k_0}^t]\}\nonumber\\ & = \mathrm{trace}\{h^t_{k_0}h_{k_0 } \sum_{i=1}^r z_iz_i^t\ } \nonumber\\ & = \sum_{i=1}^r [ z^t_ih^t_{k_0}h_{k_0 } z_i]=0 , \label{eq bjbjb}\end{aligned}\ ] ] which gives @xmath364 for @xmath365 . similarly , ( [ eq vghvh ] ) yields @xmath366 according to the given assumptions , we must have @xmath367 , @xmath368 , which again implies @xmath300 a.s .. by lemma [ lem:3.1.2 ] , it is known that under the conditions of lemma [ lem:3.1.2 ] , @xmath300 is the unique @xmath302-unobservable state , i.e. , @xmath369 . [ lem:3.1.3]uniform detectability implies @xmath0-exact detectability . for any @xmath370-unobservable state @xmath371 , by definition [ def:2.2 ] and definition [ def:3.1.3 ] , we must have @xmath372 or @xmath373 for @xmath158 ; otherwise , it will lead to a contradiction since @xmath374 under any case , the following system @xmath375 is esms , so @xmath111 is exactly detectable . [ rem:3.1.2]when system ( [ eq vcgtgh1 ] ) reduces to the deterministic time - invariant system ( [ eq vgthgh ] ) , the uniform detectability , @xmath21-exact detectability and @xmath0-exact detectability coincide with the classical detectability of linear systems @xcite . examples [ exm:3.1.4][exm:3.1.5 ] show that there is no inclusion relation between uniform detectability and @xmath11-exact detectability for some @xmath376 . we conjecture that if @xmath111 is uniformly detectable , then there is a sufficiently large @xmath376 such that @xmath111 is @xmath11-exactly detectable . corresponding to theorem [ thm:2.1 ] , we also have the following theorem for exact detectability , but its proof is very simple . [ thm:3.1.1]if @xmath111 is @xmath0-exactly detectable , then so is @xmath186 for any output feedback @xmath184 . we prove this theorem by contradiction . assume that @xmath186 is not @xmath0-exactly detectable . by definition [ def:3.1.3 ] , for system ( [ eq vcgfdef ] ) , although the measurement equation becomes @xmath377 for @xmath119 , the state equation @xmath378 is not esms . in view of @xmath377 , ( [ eq cvghvh ] ) is equivalent to @xmath379 hence , under the condition of @xmath377 for @xmath380 , if ( [ eq cvghvh ] ) is not esms , then so is ( [ eq dvbfhd ] ) , which contradicts the @xmath0-exact detectability of @xmath111 . it should be pointed out that theorem [ thm:3.1.1 ] does not hold for @xmath3-exact detectability . that is , even if @xmath111 is @xmath3-exactly detectable for @xmath324 , @xmath186 may not be so , and such a counterexample can be easily constructed . [ prop:3.1.1]if there exists a matrix sequence @xmath381 such that @xmath382 is esms , then @xmath111 is @xmath0-exactly detectable . because ( [ eq dvbfhdvgbg ] ) is esms , by proposition [ prop:2.1 ] and lemma [ lem:3.1.3 ] , @xmath383 is @xmath4-exactly detectable . by theorem [ thm:3.1.1 ] , for any matrix sequence @xmath384 , @xmath385 is also @xmath0-exactly detectable . taking @xmath386 , we obtain that @xmath111 is @xmath4-exactly detectable . thus , this proposition is shown . [ rem:3.1.3]in some previous references such as @xcite , if system ( [ eq dvbfhdvgbg ] ) is esms for some matrix sequence @xmath387 , then @xmath111 is called stochastically detectable or detectable in conditional mean @xcite . proposition [ prop:3.1.1 ] tells us that stochastic detectability implies @xmath0-exact detectability , but the converse is not true . such a counterexample can be easily constructed ; see the following example [ exm:3.1.6 ] . the @xmath388-exact detectability implies that any @xmath289-unobservable initial state @xmath314 leads to an exponentially stable trajectory for any @xmath299 . however , in the time - invariant system ( [ eq vcgdsdsf ] ) , the stochastic detectability of ( [ eq vcgdsdsf ] ) ( or @xmath389 for short ) is equivalent to that there is a constant output feedback gain matrix @xmath390 , rather than necessarily a time - varying feedback gain matrix sequence @xmath391 , such that @xmath392 is esms ; see @xcite . [ exm:3.1.6]let @xmath393 for @xmath127 , and @xmath394 by lemma [ lem:2.2 ] , for any output feedback @xmath184 , we have @xmath395 for @xmath396 , where @xmath43 is the closed - loop trajectory of @xmath397 which is not esms . so @xmath398 is not stochastically detectable . however , @xmath111 is not only @xmath16- but also @xmath399-exactly detectable , and @xmath400 is the unique @xmath401-unobservable state . [ rem:3.1.4]according to the linear system theory , for the deterministic linear time - invariant system ( [ eq vgthgh ] ) , the @xmath0- and @xmath21-exact detectability are equivalent . by the @xmath402-representation theory @xcite , for ( [ eq vcgdsdsf ] ) , the @xmath0- and @xmath403}$]-exact detectability are also equivalent . so , in what follows , system ( [ eq vcgdsdsf ] ) ( or @xmath389 ) is simply called exactly detectable . [ rem:3.1.5]in example [ exm:3.1.3 ] , @xmath329 is stochastically detectable , but it is not @xmath11-exactly detectable for any @xmath324 . in example [ exm:3.1.6 ] , @xmath329 is not stochastically detectable , but it is @xmath11-exactly detectable for @xmath404 . hence , it seems that there is no inclusion relation between stochastic detectability and @xmath11-exact detectability . at present , we do not know whether theorem [ thm:4.1.2 ] holds under exact detectability , but we are able to prove a similar result to theorem [ thm:4.1.2 ] for a periodic system , namely , in ( [ eq vcgtgh1 ] ) , @xmath405 , @xmath406 , @xmath407 . periodic systems are a class of very important time - varying systems , which have been studied by many researchers ; see @xcite . [ thm:5.3.1]assume that system ( [ eq vcgtgh1 ] ) is a periodic system with the period @xmath408 . if system ( [ eq vcgtgh1 ] ) is @xmath5-exactly detectable for any fixed @xmath324 and @xmath409 is a positive definite matrix sequence which solves gle ( [ eq vhvcxf ] ) , then the periodic system ( [ eq vcgtgh1 ] ) is esms . by periodicity , @xmath410 . select an integer @xmath411 satisfying @xmath412 . for @xmath413 , we introduce the following backward difference equation @xmath414 set @xmath415 , then associated with ( [ eq vffggjjj ] ) , we have @xmath416 = x_0^tp_0x_0-e[x_{\kappa\tau}^tp_{0}x_{\kappa\tau}]\nonumber\\ & = \sum_{i=0}^{\kappa\tau-1}e\|y_i\|^2=x_0^tp_0^{\kappa\tau-1}(0)x_0 , \label{eq cvgvgfgf}\end{aligned}\ ] ] where the last equality is derived by using the completing squares technique . we assert that @xmath417 . otherwise , there exists a nonzero @xmath56 satisfying @xmath418 due to @xmath419 . as so , by @xmath3-exact detectability , ( [ eq cvgvgfgf ] ) leads to @xmath420 where @xmath421 and @xmath315 are defined in ( [ eq vcgv ] ) . if @xmath422 is taken sufficiently large such that @xmath423 with @xmath424 being a minimal integer satisfying @xmath425 , then ( [ eq vgvhgf ] ) yields @xmath426 , which contradicts @xmath427 . if we let @xmath428 denote the solution of @xmath429 then by periodicity , @xmath430 , and @xmath431\\ & = e[x_{(n-1)\kappa\tau}^tp_{0}^{\kappa\tau-1}(0)x_{(n-1)\kappa\tau } ] \ge \varrho_0 \|x_{(n-1)\kappa\tau}\|^2 , \end{aligned}\ ] ] where @xmath432 . generally , for @xmath433 , we define @xmath434 as the solution to @xmath435 it can be shown that @xmath436 and @xmath437,\ ] ] provided that we take @xmath438 , where @xmath439 is the minimal integer satisfying @xmath440 . summarizing the above discussions , for any @xmath127 and @xmath441 , we have @xmath442 where @xmath443 with @xmath444 $ ] . the rest is similar to the proof of theorem [ thm:4.1.2 ] and thus is omitted . in theorem [ thm:5.3.1 ] , if @xmath409 is weaken as @xmath286 , then we have [ thm:5.3.2]assume that system ( [ eq vcgtgh1 ] ) is a periodic system with the period @xmath408 . if ( i ) system ( [ eq vcgtgh1 ] ) is @xmath5-exactly detectable for any fixed @xmath324 ; ( ii ) @xmath445 is a positive semi - definite matrix sequence which solves gle ( [ eq vhvcxf ] ) ; ( iii ) @xmath446 , then the periodic system ( [ eq vcgtgh1 ] ) is esms . from gle ( [ eq vhvcxf ] ) , it is easy to show ( e.g. , see theorem 3.2 in @xcite ) that @xmath447 , @xmath448 , @xmath449 . in addition , in view of @xmath450 and @xmath451 , there is a common orthogonal matrix @xmath452 such that for any @xmath127 , there hold @xmath453 , \ p^{22}_k\ge 0 , \ s^th_k^th_ks=\left [ \begin{array}{cc } 0 & 0\\ 0 & ( h^{22}_k)^t h^{22}_k \end{array } \right],\ ] ] @xmath454 , \ \ s^tg_ks=\left [ \begin{array}{cc } g^{11}_k & g^{12}_k\\ 0 & g^{22}_k \end{array } \right].\ ] ] pre- and post - multiplying @xmath455 and @xmath452 on both sides of gle ( [ eq vhvcxf ] ) gives rise to @xmath456 which is equivalent to @xmath457 set @xmath458=s^tx_k= \left [ \begin{array}{cc } s_{11 } & s_{12}\\ s_{21 } & s_{22 } \end{array } \right]^tx_k$ ] , then it follows that @xmath459 it can be easily seen that @xmath460 iff @xmath461 , for which a sufficient condition is @xmath462 . by @xmath3-exact detectability , @xmath463 is esms . to show that @xmath464 is esms , we consider the following reduced - order state - measurement equation @xmath465 obviously , system ( [ eq vhvhhss ] ) is still a periodic system and has the same period @xmath408 as ( [ eq vcgtgh1 ] ) . in the following , we show that ( [ eq vhvhhss ] ) is also @xmath466-exactly detectable . because system ( [ eq vcgtgh1 ] ) is @xmath3-exactly detectable , for any @xmath127 , @xmath467 a.s . for @xmath468 , implies that there are constants @xmath469 and @xmath470 such that @xmath471 for any @xmath472-unobservable state @xmath43 . take @xmath473 , $ ] with @xmath474 being a @xmath472-unobservable state of ( [ eq vhvhhss ] ) , i.e. , @xmath475 for @xmath476 . then @xmath477=0 $ ] for @xmath468 . hence , ( [ eq vcgwewe ] ) holds . substituting @xmath478 $ ] into ( [ eq vcgwewe ] ) yields @xmath479 so ( [ eq vhvhhss ] ) is @xmath3-exactly detectable . associated with ( [ eq vhvhhss ] ) , the gle ( [ eq cvgvgh ] ) admits a positive definite solution sequence @xmath480 . applying theorem [ thm:5.3.1 ] , the subsystem ( [ eq vhvhhss ] ) is esms . since @xmath481 has been shown to be esms , there are constants @xmath482 and @xmath483 such that @xmath484 set @xmath485 , @xmath486 , then the composite system ( [ eq sdfthtyhd ] ) is esms with @xmath487 which deduces that the periodic system ( [ eq vcgtgh1 ] ) is esms because ( [ eq vcgtgh1 ] ) and ( [ eq sdfthtyhd ] ) are equivalent . finally , we consider the linear time - invariant stochastic system ( [ eq vcgdsdsf ] ) and present a lyapunov - type theorem as a complementary result of theorem 19 @xcite . associated with ( [ eq vcgdsdsf ] ) , we introduce the linear symmetric operator @xmath488 , called the generalized lyapunov operator ( glo ) , as follows : @xmath489 moreover , for system ( [ eq vcgdsdsf ] ) , the gle ( [ eq vhvcxf ] ) becomes @xmath490 [ thm:5.1.2]suppose that @xmath491 and @xmath389 is exactly detectable . if @xmath492 is a real symmetric solution of ( [ eq vhdfvf ] ) , then @xmath493 and @xmath494 is stable , i.e. , the state trajectory of ( [ eq vcgdsdsf ] ) is asymptotically mean square stable . in order to prove theorem [ thm:5.1.2 ] , we need to cite the well - known krein - rutman theorem as follows : [ lem:5.1.1]let @xmath495 be the spectral radius of @xmath488 . then there exists a nonzero @xmath496 such that @xmath497 . because @xmath498 , the spectral radius @xmath499 . if @xmath500 , then this means that @xmath494 is stable by ( * ? ? ? * lemma 3 ) , which yields @xmath9 according to ( * ? ? ? * lemma 17 ) . if @xmath501 , then by lemma [ lem:5.1.1 ] , there exists a nonzero @xmath496 , such that @xmath502 . therefore , we have @xmath503 where @xmath504 , @xmath505 is the adjoint operator of @xmath488 , and @xmath506 . from ( [ eq vhgvh ] ) it follows that @xmath507 , which implies @xmath508 due to @xmath496 . however , according to ( * ? ? ? * theorem 8-(4 ) ) , @xmath509 together with @xmath508 , contradicts the exact detectability of @xmath389 . hence , we must have @xmath510 , and this theorem is verified . [ rem:5.1.1]following the line of theorem [ thm:5.1.2 ] , conjecture 3.1 in @xcite can also be verified . this section introduce another definition called `` exact observability '' for system ( [ eq vcgtgh1 ] ) , which is stronger than exact detectability and also coincides with the classical observability when system ( [ eq vcgtgh1 ] ) reduces to the deterministic linear time - invariant system ( [ eq vgthgh ] ) . we first give the following definitions : [ def:3.2.1]system ( [ eq vcgtgh1 ] ) is called @xmath289-exactly observable if @xmath300 is the unique @xmath289-unobservable state . similarly , system ( [ eq vcgtgh1 ] ) is called @xmath292-exactly observable if @xmath300 is the unique @xmath292-unobservable state . [ def:3.2.2]system ( [ eq vcgtgh1 ] ) ( or @xmath111 ) is said to be @xmath16-exactly observable if for any @xmath511 , system ( [ eq vcgtgh1 ] ) is @xmath317-exactly observable . if for any time @xmath511 , there exists a nonnegative integer @xmath324 such that system ( [ eq vcgtgh1 ] ) is @xmath512-exactly observable , then system ( [ eq vcgtgh1 ] ) ( or @xmath111 ) is said to be @xmath3-exactly observable . similarly , the @xmath1- and @xmath2-exact observability can be defined . combining lemmas [ lem:3.1.1][lem:3.1.2 ] together , a sufficient condition for the exact observability is presented as follows . [ thm:3.2.1]if @xmath513 for any @xmath127 , then @xmath111 is @xmath16-exactly observable . in particular , if @xmath514 for some fixed integer @xmath294 and any @xmath127 , then system ( [ eq vcgtgh1 ] ) is not only @xmath16- but also @xmath515-exactly observable . here @xmath516 is defined in lemma [ lem:2.1 ] , and @xmath517 . \label{eq 2bv}\ ] ] the next corollary follows immediately from theorem [ thm:3.2.1 ] . [ cor:3.2.1]if @xmath518 is nonsingular for @xmath127 , then system ( [ eq vcgtgh1 ] ) is @xmath519-exactly observable . by definitions [ def:3.2.1][def:3.2.2 ] , @xmath289 ( resp . @xmath292)-exact observability is stronger than @xmath289 ( resp . @xmath292)-exact detectability . likewise , @xmath16 ( resp . @xmath1 , @xmath2 , @xmath3)-exact observability is stronger than @xmath16 ( resp . @xmath1 , @xmath520 , @xmath3)-exact detectability . a necessary and sufficient condition for the @xmath3-exact observability was presented in @xcite based on the @xmath521-representation theory developed therein . below , we give another equivalent theorem based on lemma [ lem:2.1 ] . [ thm:3.2.2](i ) system ( [ eq vcgtgh1 ] ) is @xmath3-exactly observable iff for any @xmath119 , the gramian @xmath22 is a positive definite matrix . ( ii ) if system ( [ eq vcgtgh1 ] ) is @xmath1-exactly observable and @xmath286 solves the gle ( [ eq vhvcxf ] ) , then @xmath522 for any @xmath127 . we note that @xmath467 a.s . for @xmath523 , is equivalent to @xmath524 . by lemma [ lem:2.1 ] , @xmath525=0 $ ] . so system ( [ eq vcgtgh1 ] ) is exactly observable iff @xmath525=0 $ ] implies @xmath145 a.s . , which is equivalent to @xmath526 due to @xmath527 . hence , ( i ) is proved . now we prove ( ii ) by contradiction . if some @xmath528 is not strictly positive definite , then there exists a nonzero @xmath529 such that @xmath530=0 $ ] . by the @xmath1-exact observability of @xmath111 , there is @xmath294 such that system ( [ eq vcgtgh1 ] ) is @xmath292-exactly observable . since the following identity @xmath531-e[x_{s+1}^tp_{s+1}x_{s+1}]&= & \sum_{i = k}^{s}e\|y_i\|^2 \label{eq fsdfddd}\end{aligned}\ ] ] holds for any @xmath532 , it follows that @xmath533\le 0\ ] ] and accordingly @xmath467 a.s . for @xmath534 $ ] . by the @xmath292-exact observability , @xmath300 , which contradicts @xmath535 . hence , ( ii ) is proved . [ rem:3.2.1]theorem [ thm:3.2.2]-(i ) shows that the @xmath3-exact observability is weaker than the uniform observability given in @xcite , where it was proved that system ( [ eq vcgtgh1 ] ) is uniformly observable iff there are @xmath324 and @xmath536 such that @xmath537 for any @xmath119 . [ rem:3.2.2]there is no inclusion relation between uniform detectability and exact observability . for example , in example [ exm:3.1.4 ] , @xmath329 is @xmath11-exactly observable , but it is not uniformly detectable . on the other hand , in example [ exm:3.1.5 ] , @xmath329 is uniformly detectable , but it is not @xmath344-exactly observable . similar to exact detectability , we also have the following inclusion relation for exact observability : @xmath538 the following lyapunov - type theorem can be viewed as a corollary of theorem [ thm:5.3.1 ] . [ thm:4.2.1]assume that system ( [ eq vcgtgh1 ] ) is a periodic system with the period @xmath408 . if system ( [ eq vcgtgh1 ] ) is @xmath5-exactly observable for any @xmath324 and @xmath539 solves gle ( [ eq vhvcxf ] ) , then the periodic system ( [ eq vcgtgh1 ] ) is esms . by theorem [ thm:3.2.2]-(ii ) , @xmath522 for @xmath127 . because @xmath3-exact observability must be @xmath3-exact detectability , this theorem is an immediate corollary of theorem [ thm:5.3.1 ] . at the end of this paper , we give the following comments : * in @xcite , exact observability and detectability of linear stochastic time - invariant systems with markov jump were studied . how to extend various definitions of this paper to linear time - varying markov jump systems is an interesting research topic that merits further study . * following the line of @xcite that transforms the system ( [ eq vcgtgh1 ] ) into a deterministic time - varying system , it is easy to give some testing criteria for uniform detectability and observability of ( [ eq vcgtgh1 ] ) by means of the existing results on deterministic time - varying systems@xcite . in addition , applying the infinite - dimensional operator theory , the spectral criterion for stability of system ( [ eq vcgtgh1 ] ) is also a valuable research issue . * in view of remarks [ rem:3.1.3][rem:3.1.4 ] , we know that , for linear time - invariant system ( [ eq vcgdsdsf ] ) , stochastic detectability implies exact detectability . in @xcite , it was shown that exact detectability is equivalent to the so - called w - detectability " ( see ( * ? ? ? * definition 3 ) ) . a new definition called weak detectability " was introduced in @xcite , where a counter - example ( see example 15 in @xcite ) shows that w - detectability does not imply weak detectability . in particular , it was proved in @xcite that weak detectability can be derived from stochastic detectability . it is easy to prove that weak detectability implies exact detectability . in summary , we have the following inclusion relation : @xmath540 as stated in @xcite , the converse implication that whether w - detectability or exact detectability implies weak detectability is an open question . * lemmas [ lem:2.2][lem:2.1 ] are important , which will have potential applications in mean stability analysis and system synthesis . * this paper reveals some essential differences between linear time - varying and time - invariant systems . for example , for linear time - invariant system ( [ eq vcgdsdsf ] ) , exact detectability and exact observability can be uniquely defined , but they exhibit diversity for ldtv system ( [ eq vcgtgh1 ] ) . moreover , many equivalent relations in linear time - invariant system ( [ eq vcgdsdsf ] ) such as @xmath541 do not hold for ldtv system ( [ eq vcgtgh1 ] ) . this paper has introduced the new concepts on detectability and observability for ldtv stochastic systems with multiplicative noise . uniform detectability defined in this paper can be viewed as an extended version of that in @xcite . various definitions on exact detectability and observability are extensions of those in @xcite to ldtv stochastic systems . different from time - invariant systems , defining exact detectability and exact observability for the time - varying stochastic system ( [ eq vcgtgh1 ] ) is much more complicated . we have also obtained some lyapunov - type theorems under uniform detectability of ldtv systems , @xmath11-exact detectability and @xmath11-exact observability of linear discrete periodic systems . we believe that all these new concepts that have been introduced herein will play important roles in control and filtering design of ldtv systems . the first author would be grateful to his doctoral students yong zhao and yaning lin for their valuable discussions in proving lemmas [ lem:2.2][lem:2.1 ] . w. liu , f. deng , x. yan , and y. zhao , the detectability of discrete - time stochastic linear systems with markovian jump and multiplicative noises , " _ proceedings of 26th chinese control and decision conference _ , changsha , china , may 31june 2 , 2014 , pp . 17551758 . a. samir and v. dragan , robust stability and robust stabilization of a class of discrete - time time - varying linear stochastic systems , " _ siam j. control and optimization _ , vol . 53 , no . 1 , pp . 3057 , 2015 . v. v. dombrovskii and e. a. lyashenko , a linear quadratic control for discrete systems with random parameters and multiplicative noise and its application to investment portfolio optimization , " _ automat . remote control _ 64 , pp . 15581570 , 2003 . y. h. ni , w. zhang , and h. t. fang , on the observability and detectability of linear stochastic systems with markov jumps and multiplicative noise , " _ j. syst . sci . complexity _ , vol . 21 , no . 1 , pp . 102115 , 2010 . m. a. peters and p. a. iglesias , a spectral test for observability and detectability of discrete - time linear time - varying systems , " _ proceedings of the 36th conference on decision and control _ , san diego , califomia , usa , pp . 39203925 , 1997 . j. seo , d. chung , c. g. park , and j. g. lee , the robustness of controllability and observability for discrete linear time - varying systems with norm - bounded uncertainty , " _ ieee trans . 50 , pp . 10391043 , 2005 . v. ungureanu , stability , stabilizability and detectability for markov jump discrete - time linear systems with multiplicative noise in hilbert spaces , " _ optimization : a jpurnal of mathematical programming and operations research _ , vol . 63 , no . 11 , pp . 16891712 , 2014 . v. ungureanu and v. dragan . stability of discrete - time positive evolution operator on ordered banach spaces and applications , " _ j. difference equations and applications _ , vol . 19 , no . 6 , pp . 952980 , 2013 . c. w. yao , b. d. hsu , and b. s. chen , constructing gene regulatory networks for long term photosynthetic light acclimation in arabidopsis thaliana , " _ bmc bioinformatics _ , vol . 12 , no . 335 , 2011 . http://www.biomedcentral.com/1471-2105/12/335 w. zhang , h. zhang , and b. s. chen , generalized lyapunov equation approach to state - dependent stochastic stabilization / detectability criterion , " _ ieee trans . 53 , pp . 16301642 , 2008 .
the objective of this paper is to study detectability , observability and related lyapunov - type theorems of linear discrete - time time - varying stochastic systems with multiplicative noise . some new concepts such as uniform detectability , @xmath0-exact detectability ( resp . @xmath1-exact detectability , @xmath2-exact detectability , @xmath3-exact detectability ) and @xmath4-exact observability ( resp . @xmath1-exact observability , @xmath2-exact observability , @xmath5-exact observability ) are introduced , respectively , and nice properties associated with uniform detectability , exact detectability and exact observability are also obtained . moreover , some lyapunov - type theorems associated with generalized lyapunov equations and exponential stability in mean square sense are presented under uniform detectability , @xmath3-exact observability and @xmath3-exact detectability , respectively . * key words : * discrete - time time - varying stochastic systems , generalized lyapunov equations , uniform detectability , exact detectability , exact observability .
recently , it has become possible to perform single - molecule experiments . @xcite in such single - molecule experiments or small chemical systems such as cells , it is expected that ` fluctuation ' plays an important role because mean values and fluctuations ( deviations ) of observables are in the same order . @xcite it also becomes possible to measure flux distributions experimentally in such small systems . @xcite when the fluctuation is large , the traditional rate equation approach is not adequate for the study of stochastic systems because it treats only average values in large size limit . hence , the stochastic behavior of chemical reaction systems should be treated by a master equation approach . @xcite although it is difficult to obtain exact solutions of master equations in general , it is important to evaluate ( even approximately ) not only the mean values , but also the fluctuations . there is a phenomenon , known as a _ pumping _ , in which a system under a periodic perturbation causes a finite flux in a preferred direction . for example , a classical michaelis - menten ( mm ) enzymatic mechanism under a periodic perturbation has been studied experimentally and theoretically . it has been revealed that the periodic perturbation activates a pumping mode in the mm system . @xcite that is , the periodic perturbation causes a current which is not explained by a simple average of those in the strict static cases . the concept of the pump current is also related to molecular motors and brownian ratchets . @xcite . the average current can be calculated by various methods . @xcite in addition , it has been indicated that a calculation of the net current in the brownian ratchets shows a similar mathematical structure to that of the geometric phase in quantum mechanics . @xcite recently , an explicit connection between the statistics of the current flow and the berry phase interpretation has been proposed . @xcite the total current is divided into the so - called ` classical ' current and ` pump ' current . sinitsyn and nemenman have shown that the dynamical and the berry phase correspond to the classical and pump current , respectively . @xcite one of the other remarkable progresses is that it becomes possible to calculate not only the average current , but also the fluctuation about the current for an ` adiabatic ' cases . here , the ` adiabatic ' means that the oscillation of the periodic perturbation is very slow . although the average current has already been observed experimentally , @xcite any experiment for the fluctuation has not been performed . however , recent progress of experimental techniques @xcite would make it possible to observe the flux distribution for the pumping phenomena directly . detailed information for non - equilibrium systems is useful to construct the non - equilibrium statistical mechanics and non - equilibrium thermodynamics . @xcite in order to investigate such non - equilibrium systems experimentally , theoretical predictions would become a guideline for experimentalists . in the present paper , we give a calculation scheme for the statistics of the current , especially in ` non - adiabatic ' cases . the fluctuation about the average current in the non - adiabatic cases has not been calculated yet . in addition , the non - adiabatic region is of great interest because a stochastic resonance - like phenomenon has been observed experimentally in such non - adiabatic region . @xcite one of our main results is that there is no stochastic resonance - like behavior for the fluctuation about the average current . in order to calculate the statistics , we develop a calculation scheme based on the non - adiabatic geometrical phase ( aharonov - anandan phase @xcite ) . we have already proposed a method for direct evaluation of the aharonov - anandan phase by means of a perturbation calculation for _ eigenvectors _ of a floquet hamiltonian . @xcite here , we show that a perturbation calculation for _ eigenvalues _ of a floquet hamiltonian gives the total statistics directly ; the calculation scheme is simpler than that in ref . 21 . the present paper is organized as follows . in sec . ii , we explain the stochastic model for the pumping phenomenon , which has been introduced by sinitsyn and nemenman . @xcite in sec . iii , we review a geometrical phase interpretation for the pumping phenomenon . we develop an analytical treatment for the calculation of the total current with the aid of the floquet theory and the perturbation calculation in sec . section v gives concluding remarks . we mainly focus on a simple stochastic system with two absorbing states , which has been introduced in ref . the classical stochastic system is described as follows : @xmath0 \leftrightarrows [ \mathrm{c } ] \leftrightarrows [ \mathrm{r}].\end{aligned}\ ] ] the system consists of three parts . two absorbing states are denoted as @xmath1 $ ] and @xmath2 $ ] . these absorbing states may be interpreted in many ways : they correspond to substrate and product in a mm enzymatic reaction , or cellular compartments , and so on . these absorbing states exchange molecules or particles via an intermediate container @xmath3 $ ] . the container @xmath3 $ ] can contain either zero or one particle in it . when the container is filled with one particle , the particle can escape from the container by jumping into one of the two absorbing states , @xmath1 $ ] or @xmath2 $ ] . on the contrary , when the container is empty , either of the absorbing states can emit a new particle into the container . for simplicity , we here consider the following kinetic rates : @xmath4 \rightarrow [ \mathrm{c } ] : & k_1 = c_1 + r \cos ( \omega t),\\ \textrm{(ii ) } & [ \mathrm{l } ] \leftarrow [ \mathrm{c } ] : & k_{-1 } = c_{-1 } , \\ \textrm{(iii ) } & [ \mathrm{c } ] \rightarrow [ \mathrm{r } ] : & k_2 = c_2 , \\ \textrm{(iv ) } & [ \mathrm{c } ] \leftarrow [ \mathrm{r } ] : & k_{-2 } = c_{-2 } + r \sin ( \omega t ) , \end{array}\end{aligned}\ ] ] where @xmath5 and @xmath6 are positive real numbers , and @xmath7 is the amplitude of the perturbative oscillation . the above kinetic rates indicate that only two kinetic rates ( @xmath8 and @xmath9 ) oscillate with time at a frequency @xmath10 . note that the current @xmath11 at static cases ( @xmath12 ) , in which all kinetic rates are constant , is easily calculated from @xcite @xmath13 we here describe the current @xmath11 as _ classical current_. one might think that the time average of the current over a cyclic perturbation is simply given by the time average of the classical current @xmath11 , but it is not true . for example , when @xmath14 and @xmath15 , the time average of @xmath11 gives zero . however , even in the simple case , a net current is actually observed . such additional current is called a _ pump current_. the problem considered here is to calculate the _ total _ current which includes both classical and pump currents . our main goal is to calculate the net current between @xmath3 $ ] and @xmath2 $ ] in the steady state . in order to calculate the current , we use a method similar to the full counting statistics . @xcite let @xmath16 be the probability to have @xmath17 net transitions from @xmath3 $ ] into @xmath2 $ ] during time @xmath18 , where @xmath19 is the period of the rate oscillations . the probabilities of a filled and empty state of @xmath3 $ ] are denoted by @xmath20 and @xmath21 respectively , and the state of the system is defined by @xmath22.\end{aligned}\ ] ] due to the normalization condition , @xmath23 . by discussions in refs . 16 and 23 , the characteristic function of @xmath16 is given by @xmath24 where @xmath25 is the unit vector , @xmath26 stands for the time - ordering operator , and @xmath27.\end{aligned}\ ] ] in eq . , @xmath28 is called the counting field , and the derivatives of @xmath29 give cumulants of @xmath16 , e.g. , @xmath30 . hence , the problem for calculating the current is actually the evaluation of the characteristic function @xmath31 ( or @xmath29 ) . from eq . , it is easy to see that the characteristic function @xmath31 is related to a solution of the following differential equation : @xmath32 because the final state @xmath33 at time @xmath18 is formally calculated from @xmath34 \mathbf{p}(0).\end{aligned}\ ] ] using the normalization condition of the probability @xmath35 , we obtain the identity @xmath36 . hence , the total current @xmath37 and its derivative @xmath38 are given by @xmath39 although the existence of such cyclic state has not been mathematically justified in general cases , we checked numerically that the stochastic system reaches a steady cyclic state , starting from arbitrary initial conditions . note that when @xmath40 , eq . simply gives a master equation for the time evolution of the system , and hence @xmath41 due to the cyclic evolution of the system . we here use an analogy between the classical stochastic system and a quantum mechanical formulation . replacing the time evolution operator @xmath42 by @xmath43 , we obtain the following shr " odinger - like equation : @xcite @xmath44 the time evolution operator @xmath45 is constructed as @xmath46 when we take the initial state @xmath47 as a cyclic state , we have @xmath48 where @xmath49 is a phase factor . hence , the state vector @xmath50 is an eigenstate of the time evolution operator @xmath45 , and the phase factor @xmath49 is related to its eigenvalue . from this replacement , the problem of calculation of the total current and its fluctuation is replaced by the evaluation of the phase factor @xmath49 for eq . . note that the ` hamiltonian ' @xmath51 is a non - hermitian operator , which is different from the usual quantum mechanics . in order to calculate the total statistics , the floquet theory is available . @xcite the original floquet theory for the geometrical phase was applied to hermitian hamiltonian cases , @xcite and it was extended to non - hermitian cases . @xcite as shown in ref . 21 , the ` pump ' current is calculated from the _ eigenvectors _ of the floquet states . in contrast , we here shows that the ` total ' current is calculated from the _ eigenvalues _ of the floquet states directly . the relationship among the total phase , the dynamical phase , and the aharonov - anandan phase is shown in the appendix . from the floquet theorem , the non - unitary time - evolution operator @xmath45 for a periodic non - hermitian hamiltonian @xmath51 with period @xmath18 is decomposed into the floquet product form as @xmath52 . here , @xmath45 is the unique fundamental matrix satisfying @xmath53 , and the non - unitary matrix @xmath54 also has a period @xmath18 . in addition , @xmath55 is a time - independent matrix . by using the above floquet product form , the total phase @xmath49 is related to the eigenvalues of @xmath56 via @xmath57 where we define the initial state @xmath58 as the eigenvector of @xmath56 . the index @xmath59 specifies an eigenvector and its corresponding eigenvalue . equation means that @xmath60 is the eigenvalues of @xmath55 . from the above discussion , it is needed to calculate the time evolution operator @xmath45 or the matrix @xmath55 and then the evaluation of its eigenvalues should be done . however , it is difficult to obtain them in general . in order to evaluate the total phase @xmath61 , we here introduce another fundamental matrix and its floquet product form @xmath62 , in which @xmath63 is assumed to be a diagonal matrix . @xcite the fundamental matrix @xmath64 is useful for the fourier analysis . since @xmath65 and @xmath66 are both fundamental matrices of @xmath51 , there exists a constant invertible matrix @xmath67 with @xmath68 . @xcite because @xmath53 , we have @xmath69 . hence , we obtain @xmath70 and @xmath71 here , we can make identifications @xmath72 and @xmath73 . using the above identifications and @xmath74 we obtain @xmath75 where @xmath76 is the eigenvector of @xmath63 . although the matrix @xmath63 may not be taken as a diagonal matrix in some convenient bases in general cases , in our case the matrix @xmath63 can be diagonalized . @xcite hence , the diagonal elements of @xmath63 correspond to @xmath77 directly . from the above discussions , all we have to do is to calculate the eigenvalues of @xmath63 . from the floquet theory , the matrix @xmath63 is given by the eigenvalues of the floquet hamiltonian . @xcite the floquet hamiltonian @xmath78 is defined by @xmath79 where @xmath80 the @xmath81 is an orthonormal basis for the matrix representation of @xmath78 . the index @xmath59 represents a ` state part ' and the index @xmath17 merely represents a fourier component . the ` state part ' means that the hamiltonian is expressed as @xmath82 matrix and hence the eigenvector of the hamiltonian has two components ( states ) . in our case , we denote the state parts as @xmath83 or @xmath84 ( i.e. , @xmath85 ) . the @xmath81 is called as a `` floquet state . '' note that @xmath51 is non - hermitian , so that @xmath86 is not a conjugate state of @xmath81 , but the corresponding left orthonormal basis ; we add the prime on @xmath59 in order to clarify the fact . let @xmath78 have eigenvectors @xmath87 and eigenvalues @xmath88 . the matrix elements of @xmath63 are therefore given by @xcite @xmath89 in our case , the matrix form of the floquet hamiltonian @xmath78 is written as follows : @xmath90 in the bases @xmath91 . it is easy to see that when @xmath92 , we have the block diagonalized matrix , and the block diagonalized elements have a periodic behavior . the block diagonalized hamiltonian for @xmath93 is given by @xmath94 the eigenvalues and right eigenvectors of the non - perturbative hamiltonian @xmath95 are denoted by @xmath96 and @xmath97 , respectively . since @xmath78 is not hermitian , we define the left eigenvectors of @xmath95 as @xmath98 . from the eigenvalues of the block - diagonalized hamiltonian @xmath95 , we have the following total phase as the zeroth order of @xmath7 : @xmath99 , \label{eq_classical_phase}\end{aligned}\ ] ] where @xmath100 . equation indicates that the index @xmath83 is adequate . hence , we select @xmath101 as the total phase . it is easy to see that the total phase for @xmath102 case recovers the static current @xmath11 ( eq . ) adequately . for @xmath103 case , the usual perturbation theory can be used to calculate @xmath104 . there is no first order correction for the eigenvalues @xmath105 , and hence we consider the second order correction : @xmath106 where @xmath107 is the rest of the floquet hamiltonian @xmath78 , which includes the perturbative parameter @xmath7 . after the perturbation calculation , it is possible to calculate the total current @xmath37 and its fluctuation @xmath38 as follows : @xmath108 we here consider a simple case with @xmath109 . in this case , the total current and its fluctuation up to the second order of @xmath7 are explicitly given as @xmath110 from these expressions , it is easy to see the effects of the perturbation , as follows . ( eq . ) . ( b ) its fluctuation @xmath38 ( eq . ) . here , we set @xmath111 , and @xmath112 . , title="fig:",width=302 ] + ( eq . ) . ( b ) its fluctuation @xmath38 ( eq . ) . here , we set @xmath111 , and @xmath112 . , title="fig:",width=302 ] first , the perturbation induces a phenomenon like a stochastic resonance . figure [ fig_result](a ) shows the total current @xmath37 calculated from eq . . here , we set @xmath112 . the total current shows a peak at a certain frequency @xmath113 , which has been also observed experimentally . @xcite in the experiment in ref . 7 , na@xmath114 pumping mode of ( na , k)-atpase by an oscillating electric field has been studied . because different conformations of the protein have different dipole moments , an oscillating electric field could drive structural change of the protein and hence cause the modulation of kinetic parameters . the actual dependency of the kinetic parameters on the external field could be very complicated , and we assume that two of the kinetic rates depend on the oscillating field in our model . our model is a very simplified one , so that the quantitative comparison with the experimental results may be difficult . however , the qualitative behavior is the same as the experimental one : there is one peak , and it decays as @xmath115 for @xmath116 . @xcite hence , we believe that our model catches the feature of the pumping phenomenon . second , the perturbation varies the fluctuation of the total current , too . while it is possible to evaluate the total current by various analytical methods , @xcite to our knowledge , the fluctuation has not been calculated explicitly for a long time . in adiabatic cases , the fluctuation has been calculated via the berry phase interpretation . @xcite in the adiabatic cases , it has been revealed that the perturbation decreases the fluctuation of the current . from eq . , it is clear that the perturbation decreases the fluctuation of the total current even in the non - adiabatic case . the fluctuation @xmath38 is shown in fig . [ fig_result](b ) . in contrast to the current , fig . [ fig_result](b ) shows no peak : the fluctuation increases monotonically with the frequency @xmath10 , and approaches to the value of the non - perturbative case ( in this case , @xmath117 ) . finally , we comment on the relationship between the present calculation and the aharonov - anandan phase . because the second order correction vanishes when we set @xmath102 , one might consider that all the second order correction stems from the aharonov - anandan phase . however , it is not true . for the simplest case with @xmath111 , the classical current is zero , and the non - zero total current of eq . stems from the aharonov - anandan phase . in contrast , the fluctuation calculated from the aharonov - anandan phase becomes zero , so that the second order term in eq . is induced from the dynamical phase . in summary , the second order correction does not stem from only the aharonov - anandan phase , but also the dynamical phase . if one wants to know only the ` pumping ' current and its fluctuation , one more calculation for the aharonov - anandan phase is needed . in this work , we gave a unified theory for the calculation of the total current and its fluctuation in classical stochastic systems under a periodic perturbation of the kinetic rates . the formulation is based on the full counting statistics and the non - adiabatic geometrical phase . it was clarified that the total current is easily obtained by the combination of the floquet theory and a simple perturbation calculation . although it is possible to calculate the aharonov - anandan phase directly , @xcite the calculation is a little complicated . in a practical sense , the total current would be important , and as shown in this work , the total current is easily obtained only by the perturbation calculation for the eigenvalues of the floquet hamiltonian ; the calculation is easier than the direct evaluation of the aharonov - anandan phase . hence , the formulation given in the present paper gives a useful analytical method for the calculation of the current in the non - adiabatic cases . since our model is a simplified one , we can obtain the analytical expressions for the total current and its fluctuation . although it may be difficult to compare the simple model and real experiments , an important theoretical prediction is that there is no peak structure for the current fluctuation , which is different from the behavior of the average current . the simple model can predict the qualitative behavior for the average current ( one peak and the decay @xmath115 ) adequately , so that we expect that the qualitative behavior for the fluctuation of the average current obtained in the present paper will be observed in real experiments . in addition , we hope that our analytical treatments and results become a basis for further studies , especially for the connection with the non - equilibrium thermodynamics . we thank n. a. sinitsyn for useful comments , and anonymous referees for many valuable information and suggestions . it is possible to divide the total phase @xmath49 in eq . into two parts as follows : @xcite @xmath118 @xmath119 is called the dynamical phase , and @xmath120 is the aharonov - anandan phase . the dynamical phase is defined by @xcite @xmath121 where @xmath50 is a state vector of a cyclic evolved state , and @xmath122 is related to a evolved state associated with the periodic adjoint hamiltonian @xmath123 . the aharonov - anandan phase is calculated from @xcite @xmath124 it has been shown that the aharonov - anandan phase gives the pump current . @xcite that is , there are the following correspondences between the phases and currents : @xmath125 the total current is given by @xmath126 . higher cumulants are also calculated from these phases . in ref . 21 , we have developed a theory for the perturbative calculation for the eigenvectors of the floquet state , which is related to the state vector @xmath127 . in contrast , in the present paper , we clarified that the total current is directly related to the eigenvalues of the floquet hamiltonian . in practical sense , it is needed to evaluate the total current and its fluctuation , because experimental observables include these two effects . although it is possible to evaluate the pump current from the aharonov - anandan phase , @xcite it is more convenient for us to calculate the total phase directly . c. v. rao , d. m. wolf , and a. p. arkin , nature * 420 * , 231 ( 2002 ) . m. b. elowitz , a. j. levine , e. d. siggia , and p. s. swain , science * 297 * , 1183 ( 2002 ) . e. seitaridou , m. m. inamdar , r. phillips , k. ghosh , and k. dill , j. phys . b * 111 * , 2288 ( 2007 ) . r. d. astumian and b. robertson , j. chem . phys . * 91 * , 4891 ( 1989 ) . b. robertson and r. d. astumian , j. chem . * 94 * , 7414 ( 1991 ) . k. jain , r. marathe , a. chaudhuri , and a. dhar , phys . 99 * , 190601 ( 2007 ) . y. aharonov and j. anandan , phys . lett . * 58 * , 1593 ( 1987 ) . a. borm , a. mostafazadeh , h. koizumi , q. niu , and j. zwanziger , _ the geometric phase in quantum systems _ ( springer - verlag , berlin , 2003 ) . one can continue the following discussions without introducing such analogy of the quantum mechanical formulation , by the use of the fiber bundle . although the replacement of @xmath128 by @xmath51 may have no physical meanings , such quantum mechanical formulation has been developed well , and hence we here use the replacement . h. choutri , m. maamache , and s. menouar , j. korean phys . soc . * 40 * , 358 ( 2002 ) d. j. moore and g. e. stedman , j. phys . a : math . gen . * 23 * , 2049 ( 1990 ) . d. j. moore , j. phys . a : math . gen * 23 * , l665 ( 1990 ) .
we calculate a current and its fluctuation in a two - state stochastic system under a periodic perturbation . the system could be interpreted as a channel on a cell surface or a single michaelis - menten catalyzing enzyme . it has been shown that the periodic perturbation induces so - called pump current , and the pump current and its fluctuation are calculated with the aid of the geometrical phase interpretation . we give a simple calculation recipe for the statistics of the current , especially in a non - adiabatic case . the calculation scheme is based on the non - adiabatic geometrical phase interpretation . using the floquet theory , the total current and its fluctuation are calculated , and it is revealed that the average of the current shows a stochastic - resonance - like behavior . in contrast , the fluctuation of the current does not show such behavior .
the fermionic isotope of potassium @xmath0k is among the most interesting atomic species for the experimental study of quantum degenerate gases , as early discussed in @xcite . recently , the possibility of cooling a sample of fermionic @xmath0k atoms to the regime of quantum degeneracy has been demonstrated @xcite , by means of evaporative cooling in a magnetic trap . this is the first step towards further studies on a dilute fermi system , including the possibility of cooper pairing . as for this prospect , the presence of feshbach resonances in the scattering length of selected zeeman substates of @xmath0k @xcite , induced by a magnetic field , could be exploited to bring the required temperatures for a bcs - like phase transition to a range experimentally achievable . optical traps appear to be very useful tools for investigating such collisional properties of atoms at low temperature , since they can hold high - density samples in any spin state , and an homogeneous magnetic field can be applied without modification of the trapping potential @xcite . for the particular case of @xmath0k , the most interesting feshbach resonance has been predicted for two spin states which are not magnetically trappable , and therefore an optical trap would be necessary @xcite . in this paper we report for the first time on optical trapping of a sample of @xmath0k , precooled in a magneto - optical trap to sub - doppler temperatures @xcite . by using a compression phase followed by optical molasses cooling we obtain trapped samples at relatively low temperature ( 65 @xmath4k ) and high density ( 10@xmath2 @xmath3 ) . in this temperature and density regime , our optical trap appears to be an interesting environment to investigate the collisional properties of @xmath0k , possibly including the predicted feshbach resonances . we are indeed able to perform elastic collisional measurements on unpolarized sample of atoms in the ground f=9/2 state , and inelastic measurements on samples in the f=9/2 and f=7/2 hyperfine substates , which are both held in the optical trap . a comparison with theoretical predictions is made in both cases , and a good agreement is found . the optical trap is realized with linearly polarized light from a single - mode ti : sa laser , detuned to the red of both the d@xmath5 and d@xmath6 transitions of potassium , respectively at 769.9 nm , and 766.7 nm . the laser radiation is arranged in a vertical standing - wave configuration , by retroreflecting the beam , achieving a 1d optical lattice with the strongest confinement against gravity . the laser beam is weakly focused within a two - lens telescope to a waist - size @xmath7=90 @xmath4 m , with a rayleigh length @xmath8=3 cm ; the effective running wave power at the waist position is p=600 mw . a schematic of the experimental apparatus is shown in fig . [ apparatus ] . the use of a tunable ti : sa laser allows us to change in a continuous way the trap detuning from a few nanometers to about 40 nm , allowing a variety of trap conditions . for a typical wavelength of the trapping laser @xmath9=787 nm , the calculated trap depth@xcite for the ground state is @xmath10=300 @xmath4k ( 6.2 mhz ) and the scattering rate is @xmath11=30 s@xmath12 . since the trapping light is very far detuned from any transition starting from the excited 4p states , the light shift for both the excited states is positive , and it is approximately half of the trap depth . the vibrational frequencies of each lattice site are given , in the harmonic approximation , by @xmath13 in the axial and radial directions , respectively . here @xmath14 is the mass of potassium . the expected values of the radial and axial trap frequencies for the conditions above are respectively @xmath15870 hz and @xmath15450 khz . the power of the trapping laser can be controlled with a fast aom ( with a fall time of 300 ns ) ; a polarizer cube after the aom provides the required linear polarization . the pump laser is a low - noise doubled nd : yvo , and no active stabilization or tuning of the ti : sa laser frequency is used , to avoid amplitude noise which can possibly heat up the trapped atoms @xcite . we load the optical trap directly from a vapor cell magneto - optical trap ( mot ) of @xmath0k as described in @xcite . we note the typical cooling and repumping light intensities of 36 mw/@xmath16 and 4 mw/@xmath16 respectively , and the detuning of the cooling light of -20 mhz from the cycling f=9/2 @xmath17 f@xmath18=11/2 transition ( the natural linewidth and saturation intensities of @xmath0k are @[email protected] mhz and @xmath21=1.8 mw/@xmath16 ) . in the mot we collect in 4 s about @xmath22 atoms at a peak density of @xmath23 @xmath3 and at a temperature of 60 @xmath4k . the dipole trap beam , which is passing approximately through the center of the mot , is left on during the mot loading , with no detectable effect on the atom number or temperature . to load the optical lattice from the mot we use the following procedure : the trapped atoms are compressed for 30 ms by ramping the quadrupole gradient of the mot from the normal value of 18 gauss / cm up to 44 gauss / cm , while the intensity of the cooling and repumping light are abruptly reduced to 20 mw/@xmath16 and 0.06 mw/@xmath16 respectively . moreover , the detuning of the cooling light is reduced to -6 mhz to improve the compression effect . the mot quadrupole field is then switched off in 1 ms , and the detuning of the cooling light is increased to -20 mhz for other 2 ms , providing the usual optical molasses cooling . then the mot beams are turned off by switching off the aoms in a few @xmath4s . the repumping beams are turned off 1 ms later than the cooling ones , to pump the atoms in the lower hyperfine ground state , f=9/2 . for a complete shielding of the optical trap from the resonant stray light of the mot beams , which could cause undesired optical excitation and pumping , we also block the light with a mechanical shutter within 3 ms from the switch - off of the aoms . using this procedure , about 5% of the atoms in the mot can be loaded in the optical trap , corresponding to a few 10@xmath24 atoms . the atoms in the optical trap are probed by turning on again the mot beams at half the normal intensity and in resonance with the cycling transition , while the repumper is in resonance with the f=7/2 @xmath17 f@xmath18=9/2 . the fluorescence is detected by a photomultiplier with the help of a lens and of an optical fiber with a core diameter of approximately 0.5 mm , which provides good spatial selectivity to discard the light scattered by the windows of the vacuum cell , and allows detection of as few as 10@xmath25 atoms . a typical measurement of the lifetime of the optical trap is shown in fig . [ decayg ] . after about 0.3 s of storage , the decay curve of the atoms pumped in the ground f=9/2 state is a single exponential , with a 1/e lifetime of @xmath26=1.4(3 ) s. we expect the main source of loss in this regime to be collisions with background gas , the trap lifetime being roughly one third of the mot loading time . the faster decay at shorter times could possibly be due to losses connected to an evaporative cooling process , which we will consider below . alternatively , the density and the absolute number of trapped atoms can be measured by absorption imaging , using a pulse of light derived from the mot laser lasting 200 @xmath4s , at half the saturation intensity and in resonance with the cycling transition . to compensate for the light shift due to the optical trap , the frequency is blue shifted by approximately 1.5@xmath27 from the unperturbed resonance . the beam is passing through the optical trap in the horizontal plane and is imaged on a ccd camera with a two - lens system . the distribution of atoms in the optical trap , as shown in fig . [ ccd ] , is gaussian in the radial direction , with a typical fwhm of 70 @xmath4 m . also the axial distribution is generally gaussian , and the typical fwhm of 500 @xmath4 m indicates that approximately 1200 lattice sites , spaced by 395 nm , are occupied . the mean number density is determined by averaging the optical density over the trap section and scaling by the ratio of the lattice spacing @xmath28 to the the mean axial extension @xmath29 of an individual lattice site . such mean axial extension is estimated from the temperature and axial trap frequency as @xmath30 and is of the order of 100 nm . the typical mean density of atoms after 100 ms of storage is 7@xmath3110@xmath32 @xmath3 , and decays proportionally to the number of atoms for increasing storage times . the radial temperature is measured by detecting the radial extension of the atomic cloud after 1 ms of expansion , following a sudden switch - off of the optical trap . alternatively it is possible to estimate the temperature directly from the radial extensions of the atoms in the trap ; this estimation is affected by a larger uncertainty , since it relies on the knowledge of the radial trap frequency . the axial temperature is extracted from the temporal width of a time - of - flight ( tof ) signal , as detected by the absorption of a 1-cm wide sheet of light placed 1.5 cm below the trapping region , in resonance with the cycling transition . the tof beam , generated by an extended - cavity diode laser , is doubly passing through the vacuum cell , and the absorption signal is detected with a frequency modulation technique . the typical tof signal , as shown in fig . [ tof ] , indicates the presence of two different velocity distributions , corresponding to temperatures around 70 @xmath4k for the broad peak and around 4 @xmath4k for the narrow one . we think that the colder component is due to a small fraction of the trapped atoms which are adiabatically cooled during the release from the optical trap . indeed , the aom which controls the optical trap is capable of cutting only 95% of the light in less than 300 ns , while the residual tail is extinguished in about 5 @xmath4s . since the latter switch - off time is longer than @xmath33 , the atoms in the bottom of the trap are further cooled down . we have verified this conjecture by intentionally slowing the main switch - off time of the aom down to 5 @xmath4s , and observing an adiabatic cooling of hotter component to 8 @xmath4k . we expect to be possible to eliminate such residual adiabatic cooling effect by using the aom in a double - pass configuration , but since this would reduce by about 30% the available laser power , we prefer to fit the tof signal with a two - component curve to accurately extract the temperature of the majority of the atoms . the temperature measurement described above can be performed only after a minimum storage time of 80 ms , to allow for the separation of the tof signal due to the atoms in the optical trap from the signal due to the atoms in the mot which have not been loaded in the trap . we have verified that after this time interval the axial and radial temperature are identical within the uncertainty . to look for any relationship between the temperatures of the atoms in the mot and of those in the optical trap , we have performed a series of temperature measurements after 100 ms of storage in optical traps of different depths . the result is shown in fig . [ loading ] : the temperature appears to scale almost linearly with the trap depth , and for deep traps we measure a ratio @[email protected] . note that for very deep traps such a temperature can be much higher than the mot one , which is about 70 @xmath4k after the compression phase . this behavior , which has been reported also in @xcite , seems to indicate that the temperature of the atoms in the optical trap is determined by the process of loading and optical trapping itself , and not by the original temperature in the mot . we note also that the number of atoms loaded in the trap is observed to decrease almost linearly with the trap depth , with the density decreasing even faster due to the accompanying reduction in the spring constant of the trap . as a result , the optimal density conditions for performing collisional measurements are found only at large trap depths , where the temperature of the trapped atoms is comparable to that of the mot . a detailed description of these observations goes beyond the scope of this paper , and will therefore be presented elsewhere . we have also observed that the temperature of the atoms in the optical trap remains almost constant with time , although the calculated scattering rate would lead to a constant heating with a rate as large as 100 @xmath4ks@xmath12 for the tightest trap investigated . also this result is suggesting the presence of an evaporative cooling effect , which counteracts the heating due to photon scattering . the effective trap depth and spring constants are measured with the help of a parametric heating technique @xcite . if the power of the trap light is sinusoidally modulated by the aom , a strong heating of the atoms , accompanied by trap losses , arises whenever the modulation frequency matches twice the trap oscillation frequency or a subharmonic thereof . by monitoring the trap population after @xmath35100 ms of parametric heating it is therefore possible to measure both radial and axial trap frequency with an uncertainty of the order of 10% . a typical measurement of the trap frequencies is shown in fig . [ parametric ] , for a trap depth of 180 @xmath4k . note the presence of a strong resonance also at four times the axial frequency , which can be explained only by considering the anharmonicity of the optical trap in the axial direction , due to its sinusoidal shape @xcite . to characterize the elastic collisions we use a technique which has already proved useful in this kind of optical trap@xcite , which is based on the detection of the rethermalization of the trapped atoms after the excitation of one of the degrees of freedom . in particular , we parametrically excite the axial vibrational mode of the trap by applying a 2% modulation to the laser power for about 5 ms , and measure the subsequent rapid increase of the axial temperature , followed by an exponential decay due to thermalization of the axial mode with the radial ones , as mediated by elastic collisions . in order to excite uniformly the atoms distributed over the vibrational levels in the anharmonic trap , we sweep the modulation frequency roughly between @xmath36 and 2@xmath36 during the 5 ms heating phase . indeed , if only a single frequency is used , we observe a large loss of atoms from the trap , with no larger increase of the temperature , as if only the population of a few levels were excited . moreover , modulation amplitudes larger than 2% cause quite a large loss of atoms , which are detected by the tof beam together with those released from the trap , therefore altering the temperature measurements . the decay curve for the conditions @xmath37= 65@xmath4k , @xmath38=7.5@xmath3110@xmath32 @xmath3 , is shown in fig . [ decay ] ; each data point is the average of two series of about 20 tof measurements , and the original temperature was 67 @xmath4k . the fit to an exponential decay yelds @xmath39=10(2 ) s. we have checked that the decay time is halved , within the uncertainty , when the atomic density in the optical trap is halved , to verify that the rethermalization is actually mediated by elastic collisions , and not simply by cross - dimensional anharmonic effects . we compare the observed collision rates with the results of a full close - coupling numerical calculation with the standard hamiltonian including hyperfine structure . the model is parametrized by the long range dispersion coefficients of the atom - atom interaction potential and by single channel singlet and triplet scattering lengths . we adopt the high precision value @xmath403897au from ref.@xcite for the van der waals coefficient , and choose the range of possible scattering lengths by scaling to @xmath0k the results obtained for @xmath41k from an analysis of the @xmath42 @xcite molecular state photoassociation spectroscopy . similar but less restricted values are also presented in @xcite from an analysis of the @xmath43 spectrum . we obtain in this way the limits 100 @xmath44110a@xmath45 and 150 @xmath46250a@xmath45 . it is consistent with the experimental errors in the present measurements to fix the singlet scattering length to say @xmath47 and leave only the triplet as a free parameter . from our calculation and from the results of an analogous experiment performed in a magnetic trap @xcite , we expect that both @xmath48- and @xmath49-wave collisions are contributing to the cross sections at @xmath37=65 @xmath4k for @xmath0k . indeed , a @xmath49-wave shape resonance close to threshold enhances the @xmath49-wave contribution at low temperatures @xcite . since our sample is unpolarized we calculate spherical @xmath48 and @xmath49 cross sections @xmath50 , and estimate the thermalization rate as @xmath51 where @xmath52 is the average number of collisions necessary for the rethermalization , @xmath38 is the atomic density , and @xmath53 is the rms relative velocity between two colliding atoms . note that we have implicitly assumed an identical distribution of all possible spin states in f=9/2 over the optical trap , since the trap potential is independent of the magnetic moment of the atoms . we adopt for the parameters @xmath52 the values @xmath54 and @xmath55 quoted in @xcite . we find the best agreement with the measured rate at the lowest allowed value of the triplet scattering length ( @xmath56=150@xmath57 ) , _ i.e. _ when the @xmath49-wave resonance is closer to threshold . in this situation the average thermal rates are @xmath58=@xmath59cm@xmath25/s and @xmath61=@xmath62cm@xmath25/s and the characteristic time for rethermalization for the given density in the trap is then about @xmath39=20 ms . it is interesting to notice that a comparison between the photoassociation results in @xcite and @xcite seems to privilege lower values of @xmath56 in agreement with what we find here . from the calculated elastic collision rate 1/@xmath63 , with @xmath50 the total cross section , is also possible to estimate the typical speed of the thermalization due to evaporative cooling processes . using the formalism presented in @xcite , the ratio of the evaporation to the elastic collision time constant is @xmath64 where @xmath65=@xmath27/@xmath66 . from the measured value of @xmath67 ( see fig . [ loading ] ) , the timescale for the thermalization due to evaporation is therefore expected to be a few hundreds of ms . although this process can be neglected with regard to the cross - dimensional collisional processes reported above , which are proceeding at a much faster rate , we speculate that it could play an important role in the observed long term stability of the temperature of the atoms in the optical trap . the inelastic two - body collisions between atoms in the f=7/2 state , and between atoms in the f=9/2 and f=7/2 states , can be detected in a relatively easy way through trap losses , since the hyperfine energy of 1.285 ghz released under the transition f=7/2@xmath179/2 is much larger than the trap depth . we can vary the population of the upper hyperfine state f=7/2 by means of an optical pumping beam , derived from the mot laser , in resonance with the f=9/2@xmath17f@xmath18=7/2 transition , and controllable in intensity with an aom . by shining the beam on the trapped atoms with an intensity of about 10 @xmath4w/@xmath16 , and for a time interval variable between 0.1 and 2 ms , we transfer up to 95% of the atoms from the ground to the upper state . the relative populations of the two hyperfine states after the pumping pulse are measured by monitoring the tof signal , to which only the atoms in f=9/2 contribute . we have verified that the typical time scale for hyperfine optical pumping due to inelastic raman scattering of the trap photons is much longer than 1 s , and therefore such effect is negligible . the optical pumping pulse is applied after about 100 ms from the loading of the dipole trap , to let the trapped atoms equilibrate , and then the decay of the number of atoms in the trap is measured with the usual procedure . in fig . [ hyperfine ] the decay of the trapped atoms is shown for the cases of 95% and 85% of the total population pumped into the f=7/2 state . in this particular measurement the temperature of the atoms was @xmath37=50 @xmath4k . we fit the experimental data by assuming a constant volume regime , which in an harmonic trap is equivalent to the assumption of a constant temperature . in this regime , the density is proportional to the number of atoms , and the decay curves are the solutions of the coupled differential equations for the densities of the populations in both hyperfine states @xmath68 here the suffixes 1 and 2 refer to the levels f=9/2 and f=7/2 , respectively , @xmath69 is the rate of linear trap loss , and @xmath70 is the spherical rate of two - body inelastic collisions between atoms in states @xmath71 and @xmath72 at temperature @xmath37 . to account for the loss of both atoms in the pair for each collision event , factors of 2 are added for collisions of atoms belonging to the same hyperfine level . the constant @xmath69 can be obtained from the tails of the decay curve , where the binary collision rate is negligible . the fit yields @xmath69=([email protected] ) s@xmath12 . we keep this value fixed for the fits where we have 95% , 83% and 66% of the atoms in f=7/2 . we assume a 30% uncertainty in the atomic density and obtain best estimates of the decay rates from the individual curves . next we combine the three measurements to obtain @xmath74=3.2@xmath3110@xmath75 cm@xmath25/s and @xmath76=3.8@xmath3110@xmath77 cm@xmath25/s with an uncertainty of the order of 50% . from numerical calculations we find thermally averaged spherical rates of @xmath74=4.4@xmath3110@xmath78 cm@xmath25/s and @xmath76=3.5@xmath3110@xmath77 cm@xmath25/s for @xmath37=50@xmath4k , in agreement with the measured rate . we obtain these values at the lower edge @xmath79=150@xmath57 of the initial confidence range . the agreement holds until the value @xmath80 , where the loss coefficient @xmath76 drops to values of the order of @xmath81 cm@xmath25/s . in all this range @xmath49-wave losses provide a significant contribution to both @xmath74 and @xmath76 , and the agreement with the measured values represents therefore further evidence of the presence of a low energy @xmath49-wave resonance . it is interesting to notice that the rate coefficient @xmath82 rapidly increases above @xmath79=180@xmath57 , reaching values larger than 10@xmath78 cm@xmath25/s for @xmath83 , in disagreement with the measured value . this seem to rule out the higher range of @xmath79 values , as already indicated by the large observed rethermalization rate . we also note that inelastic collisions can not be responsible for the faster decay of the atom number at short trapping times ( see fig . [ decayg ] ) , even assuming a residual 5% on the f=7/2 state after the normal loading of the optical trap , due to the relatively small value of @xmath76 . we have reported on loading of fermionic @xmath0k atoms in a 1d optical lattice from a single mot . the 1.4-s lifetime of the optical trap is large enough to allow the observation of elastic and inelastic collisional processes , due to the high density of the atomic sample in the trap . indeed , the typical time scale for collisional processes is a few tens of ms , at a density of 7@xmath3110@xmath32 @xmath3 and at a temperature of 65 @xmath4k . furthermore , optical pumping and heating effects arising from the scattering of trap photons do not seem to play a significant role on such time scale . the experimental collisional rates for unpolarized samples we determine are quite consistent with the theoretical expectations . further studies of inelastic and elastic collisions in spin - polarized samples in our optical trap could lead to a characterization of the collisional properties of fermionic potassium at low temperature . as already noted , particularly interesting is the possibility of searching for the occurrence of @xmath48-wave feshbach resonances in the collisions between zeeman sublevels of the ground state . for such purpose , a temperature lower than the present 65 @xmath4k would be preferred , to reduce the @xmath49-wave contribution , and to avoid temperature broadening of the resonance . although the temperature can be lowered by loading the atoms in very shallow traps , the achieved density is not large enough to perform accurate collisional measurements , and therefore methods for reducing the temperature of a dense sample are needed . in particular , we note that the axial degree of freedom of our trap is in the lamb - dicke regime ( @xmath84 ) , allowing for the application of fast degenerate raman sideband cooling @xcite , which would allow for an efficient reduction of the temperature without atom losses . we thank d. lau for his partecipation to the early phases of the experiment , and laboratorio di fisica della scuola normale superiore for loaning the ti : sa laser for the optical trap . we also acknowledge contributions by p. hannaford and n. poli , and useful discussions with v. vuletic and f. s. cataliotti . one of us ( a. s. ) gratefully thanks e. tiesinga for helpful discussions , and the nist group for the ground state collision code . this work was supported by the european community council ( ecc ) under the contract hprict1999 - 00111 , and by murst under a prin program ( progetto di ricerca di interesse nazionale ) . w. j. was funded by the nato - cnr senior guest fellowship programme 1998 . 99 f. s. cataliotti , e. a. cornell , c. fort , m. inguscio , f. marin , m. prevedelli , l. ricci , and g. m. tino , phys . rev . a * 57 * , 1136 ( 1998 ) . b. demarco and d. s. jin , science * 258 * , 1703 ( 1999 ) . j. bohn , phys . rev . a * 61 * , 053409 ( 2000 ) . v. vuletic , c. ching , a.j . kerman , and s. chu , phys . lett . * 82 * , 1406 ( 1999 ) . g. modugno , c. benko , p. hannaford , g. roati , and m. inguscio , phys . a * 60 * , r3373 ( 1999 ) . see for example : r. grimm , m. weidemller , and y. b. ovchinikov , adv . at . . phys . * 42 * 95 ( 2000 ) . for a discussion of the heating of the trapped atoms due to noise in the trap laser , see t. a. savard , k. m. ohara , and j. e. thomas phys . a * 56 * , r1095 ( 1997 ) . s. j. m. kuppens , k. l. corwin , k. w. miller , t. e. chupp , and c. e. wieman , phys . rev . a * 62 * 013496 - 1 ( 2000 ) . s. friebel , c. dandrea , j. walz , m. weitz , and t. w. hnsch , phys . rev . a * 57 * , r20 ( 1998 ) . for a discussion of the classical case of parametric heating of an anharmonic oscillator see l. d. landau and e. m. lifschitz , _ mechanics _ ( pergamon , oxford , 1976 ) , pp . 80 - 84 . a. derevianko , w. r. johnson , m. s. safronova , and j. f. babb , phys . lett * 82 * , 3589 ( 1999 ) . j. p. burke , jr . , c. h. greene , j. l. bohn , h. wang , p. l. gould , and w. c. stwalley , phys . rev . a * 60 * , 4417 ( 1999 ) . c. j. williams , e. tiesinga , p. s. julienne , h.wang , w.c . stwalley , and p. l. gould , phys . a * 60 * , 4427 ( 1999 ) . b. demarco et al . , phys . lett . * 82 * , 4208 ( 1999 ) . j. bohn et . a * 59 * 3660 ( 1999 ) w. ketterle and n. j. van druten , adv . at . * 37 * 181 ( 1996 ) . v. vuletic , c. chin , a. j. kerman , and s. chu , phys . lett.*81 * , 5768 ( 1998 ) .
we report on trapping of fermionic @xmath0k atoms in a red - detuned standing - wave optical trap , loaded from a magneto - optical trap . typically , 10@xmath1 atoms are loaded at a density of 10@xmath2 @xmath3 and a temperature of 65 @xmath4k , and trapped for more than 1 s. the optical trap appears to be the proper environment for performing collisional measurements on the cold atomic sample . in particular we measure the elastic collisional rate by detecting the rethermalization following an intentional parametric heating of the atomic sample . we also measure the inelastic two - body collisional rates for unpolarized atoms in the ground hyperfine states , through detection of trap losses .
recent progress in synthesizing materials that contain ferromagnetic chains@xcite provides new opportunities to study the quantum dynamics of atomic - size domain walls ( dw ) . on the atomic level , a dw is a structure that is stable with respect to ( quantum ) fluctuations , separating two regions with opposite magnetization . such a structure was observed in the one - dimensional cocl@xmath2h@xmath3o chaintorrance69,nicoli74 . in an earlier paper@xcite , we studied the propagation of spin waves in ferromagnetic quantum spin chains that support dws . we demonstrated that dws are very stable against perturbations , and that the longitudinal component of the spin wave speeds up when it passes through a dw while the transverse component is almost completely reflected . in this paper , we focus on the dynamic stability of the dw in the heisenberg - ising ferromagnetic chain . it is known that the ground state of this model in the subspace of total magnetization zero supports dw structures@xcite . however , if we let the system evolve in time from an initial state with a dw structure and this initial state is not an eigenstate , it must contain some excited states . therefore , the question whether the dw structure will survive in the stationary ( long - time ) regime is nontrivial . the question how the dw structure dynamically survives in the stationary ( long - time ) region is an interesting problem . in particular , we focus on the stability of the dw with respect to the dynamical ( quantum ) fluctuations as we approach the quantum critical point ( from heisenberg - ising like to heisenberg ) . we show that the critical quantum dynamics of dws can be described well in terms of conventional power laws . the behavior of quantum systems at or near a quantum critical point is of contemporary interestsachdev99 . we also show that the dw profiles rapidly become very stable as we move away from the quantum critical point . the hamiltonian of the system is given bymike91,cloizeaux66,gochev77,gochev83,matt81 @xmath4 where @xmath5 indicates the total number of spins in the spin chain , and the exchange integrals @xmath6 and @xmath7 determine the strength of the interaction between the @xmath8 , @xmath9 and @xmath10 components of spin @xmath0 operators @xmath11 . here we only consider the system with the ferromagnetic ( @xmath12 ) nearest exchange interaction . it is well known that @xmath13 is a quantum critical point of the hamiltonian in eq . ( [ h1 ] ) , that is , the analytical expressions of the ground state energy for @xmath14 and @xmath15 are different and singular at the points @xmath16cloizeaux66 . in ref.@xcite gochev constructed a stable state with dw structure in both the classical and quantum treatments of the hamiltonian ( [ h1 ] ) . in the classical treatment , gochev replaces the spin operators in eq . ( [ h1 ] ) by classical vectors of length @xmath17 @xmath18and then uses the conditions @xmath19 and @xmath20 to find the ground state . in the ground state , the magnetization per site is given by@xcite @xmath21where @xmath22 , \label{sigma1}\]]@xmath23 is an arbitrary constant , and @xmath24 is a constant fixing the position of the dw . the corresponding energy is @xmath25 in the quantum mechanical treatment , gochev first constructs the eigenfunction of a bound state of @xmath26 magnons@xcite @xmath27where @xmath28@xmath29@xmath30and the corresponding energy is given by@xcite @xmath31then he demonstrated that for the infinite chain , the linear superposition @xmath32 \right\ } \left\vert \psi _ { n_{0}+i}\right\rangle , \label{magnon2}\]]where @xmath33@xmath34 ^{2}\right\ } , \]]is the quantum analog of the classical domain wall , in which @xmath35 are given in eq . ( [ szsx ] ) , and the energy coincides with eq . ( [ edw ] ) . gochev s work confirmed the existence of the dw structures in the one - dimensional ferromagnetic quantum spin @xmath0 chain . in the infinite chain , the exact quantum analog of classical dw is represented by @xmath36 . in the finite chain , the dw structure exists as a bound @xmath26-magnon state @xmath37 . the main difference between these two states is the distribution of magnetization in the xy plane . in the infinite chain , the change of the magnetization occurs in three dimensions , according to eq . ( [ szsx ] ) , but in the finite chain @xmath38 for all spins . now we consider @xmath39 of the bound state @xmath40 in the case that the number of flipped spin is half of the total spins , i.e. , @xmath41 and @xmath5 is an even number . even though the formal expression for @xmath40 is known , the expression for @xmath39 in this state ( for finite and infinite chains ) is not known . for finite @xmath5 , the ground state in the subspace of total magnetization @xmath42 can , in principle , be calculated from eq . ( [ magnon ] ) . however , this requires a numerical procedure and we loose the attractive features of the analytical approach . indeed , it is more efficient to use a numerical method and compute directly the ground state in the subspace of total magnetization @xmath42 . in fig . figure0 , we show some representative results as obtained by the power method@xcite for a chain of @xmath43 spins . in all cases , the domain wall is well - defined . obviously , because we are considering the system in the ground state , the magnetization profile will not change during the time evolution . to inject a dw in the spin chain , we take the state @xmath44 with the left half of the spins up and the other half down as the initial state ( see fig . [ figure1](a ) for @xmath43 ) . the state @xmath44 corresponds to the state with the largest weight in the bound state @xmath45 with @xmath41 , because @xmath46 reaches the maximum if @xmath47 for all @xmath48 ( note @xmath49 ) . it is clear that @xmath44 is not an eigenstate of the hamiltonian in eq . ( [ h1 ] ) . the energy of @xmath50 , relative to the ferromagnetic ground state , is @xmath51 , and its spread @xmath52 . in table [ tablea ] , we list some representative values of the energy in the initial state ( see fig . [ figure1](a ) ) and in the ground state of subspace @xmath42 ( see fig . [ figure0 ] ) . a priori , there is no reason why the dw of the initial state @xmath44 should relax to a dw profile that is dynamically stable . for @xmath53 , the difference between energy of the initial state and the ground state energies for @xmath54 is relatively large and the relative spread in energy ( @xmath55 ) is large also , suggesting that near the quantum critical point , the initial state may contain a significant amount of excited states . therefore , it is not evident that a dw structure will survive in the long - time regime . in fact , from the numbers in table i , one can not predict whether or not the dw will be stable . for instance , for @xmath56 and @xmath57 , the dw is not dynamically stable whereas for @xmath58 it is stable but the energies ( see first line in table i ) do not give a clue as to why this should be the case . on the other hand , by solving the time dependent schrdinger equation ( tdse ) , it is easy to see if the dw is dynamically stable or not . .the energy @xmath59 of the initial state @xmath60 ( see fig . [ figure1](a ) ) and the ground state @xmath61 in the @xmath42 subspace , both relative to the ground state energy of the ferromagnet . [ cols="^,^,^,^,^,^",options="header " , ] next we introduce a definition of the dw width . from eq . ( [ sznfit ] ) , we can find @xmath62 and @xmath63 which satisfy @xmath64that is , when @xmath65 equals half of its maximum value ( @xmath0 ) . here @xmath62 and @xmath63 are not necessarily integer numbers . now we can define the dw width @xmath66 as the distance between @xmath62 and @xmath63 : @xmath67 clearly , the width of the dw becomes ill - defined if it approaches the size of the chain . on the other hand , the computational resources ( mainly memory ) , required to solve the tdse , grow exponentially with the number of spins in the chain . these two factors severely limit the minimum difference between @xmath68 and the quantum critical point ( @xmath69 ) that yields meaningful results for the width of the dw . indeed , for fixed @xmath5 , @xmath70 has to be larger than the `` effective '' critical value for the finite system in order for the dw width to be smaller than the system size . although the system sizes that are amenable to numerical simulation are rather small for present - day `` classical statistical mechanics '' standards , it is nevertheless possible to extract from these simulations useful information about the quantum critical behavior of the dynamically stable dw . in fig . [ figure3 ] , we plot @xmath66 as a function of @xmath70 ( @xmath71 ) . by trial and error , we find that all the data can be fitted very well by the function @xmath72 ^{1/2}\right\ } } + b_{n } , \label{w1}\]]where @xmath73 , @xmath74 and @xmath75 are fitting parameters . as shown in fig . [ fitdifn ] , all the data for @xmath76 , @xmath77 , @xmath78 , @xmath79 and @xmath80 fit very well to eq . ( [ w1 ] ) . the results of these fits are collected in table [ table2 ] . to analyze the finite - size dependence in more detail , we adopt the standard finite - size scaling hypothesis@xcite . we assume that in the infinite system , the dw width plays the role of the correlation length , that is , we assume that@xmath81where @xmath82 is a critical exponent . finite - size scaling predicts that the effective critical value @xmath83 where @xmath73 is proportional to @xmath84 . taking @xmath85 , fig . [ fitdelta0 ] shows that @xmath86 converges to one as @xmath5 increases . as a check on the fitting procedure , we apply it to the data obtained by solving for the ground state in the @xmath42 subspace . in view of eq . ( [ sz1 ] ) and ( [ sznfit ] ) , we may expect that eq . ( [ w1 ] ) fits the data very well and , as shown in fig . [ widthpm20 ] , this is indeed the case . if we fit the data to @xmath87without assuming a priori value @xmath88 , we find that @xmath88 depends on the range of @xmath68 that was used in the fit , as shown in fig . [ exponent ] . remarkably , we find that @xmath89 if we fit the data for a large range of @xmath68 s and that @xmath88 approaches @xmath0 if we restrict the value of @xmath68 to the vicinity of the critical point . as a function of @xmath90 in a spin chain of @xmath43 spins . the exponent @xmath88 obtained by fitting the dw width to eq . ( [ w3 ] ) , with @xmath91 as obtained from the fit shown in fig . [ fitdelta0 ] , for @xmath68 in the range @xmath92 $ ] . , to describe the stability of the dw structure , we introduce @xmath93 ( @xmath94 ) : @xmath95 ^{2}}-\overline{s_{n}^{z}\left ( t_{1},t_{2};\delta\right ) } ^{2 } } , \label{delta1}\ ] ] where @xmath96 ^{2 } } \equiv\frac{\int_{t_{1}}^{t_{2}}\left\langle s_{n}^{z}\left ( t\right ) \right\rangle ^{2}dt}{t_{2}-t_{1}}.\ ] ] in order to show the physical meaning of @xmath97 , we write @xmath98 as @xmath99 where @xmath100 is a constant and @xmath101 is a time - dependent term . ( [ delta1 ] ) becomes @xmath102 ^{2}\right\ } ^{1/2}. \label{delta2}\ ] ] it is clear that if @xmath98 is a constant in the time interval @xmath103 $ ] , then @xmath104 . in general , since the initial state is not an eigenstate of the hamiltonian eq . ( [ h1 ] ) , the magnetization of each spin will fluctuate and @xmath105 . if , after long time , the system relaxes to a stationary state that contains a dw , the magnetization of each spin will fluctuate around its stationary value @xmath100 . the fluctuations are given by @xmath101 . if @xmath106 is large , the difference between the actual magnetization profile at time @xmath107 and the stationary profile @xmath100 may be large . from eq . ( [ delta2 ] ) , it is clear that @xmath108 is a measure of the deviation of @xmath109 from its stationary value @xmath100 , averaged over the time interval @xmath103 $ ] . thus , @xmath108 gives direct information about the dynamics stability of the dw . figure [ figure4 ] shows the distribution @xmath93 for different values of @xmath68 . we only show some typical results , as in fig . [ figure2 ] . as expected , the distribution of @xmath108 is symmetric about the centre of the spin chain ( @xmath110 ) . we first consider how @xmath93 changes with @xmath68 for fixed @xmath111 . from fig.[figure4 ] , we conclude : \1 ) for the spins which are not located at the dw centre , i.e. , @xmath112 , @xmath93 decreases if @xmath68 becomes larger . this means that the quantum fluctuations of these spins become smaller if we increase the value of @xmath68 . this is reasonable because with increasing @xmath68 , the initial state approaches an eigenstate of the hamiltonian for which @xmath113 ( ising limit ) . \2 ) for the spins at the dw centre , i.e. , @xmath114 , when @xmath68 becomes larger and larger , @xmath93 first increases and then decreases . qualitatively , this can be understood in the following way . when @xmath68 is close to @xmath115 , the magnetization at the dw centre disappears very fast and remains zero . however , if @xmath116 , the magnetization at the dw centre will retain its initial direction , hence the behavior of the spin at the dw centre will qualitatively change as @xmath68 moves away from the critical point @xmath117 . in fig . [ figure5 ] , we plot @xmath118 ( @xmath119 ) as a function of @xmath68 . it is clear that @xmath118 first increases as @xmath68 increases , reaches its maximum at @xmath120 , and then decreases as @xmath68 becomes larger . now we consider the @xmath111-dependence of @xmath93 for fixed @xmath68 . since @xmath93 is a symmetric function of @xmath111 , we may consider only one side of the whole chain , e.g. , the spins with @xmath121 . from fig.[figure4 ] , according to the value of @xmath68 , there are three different regions : \1 ) @xmath122 : starting from the boundary ( @xmath123 ) , @xmath124 first decreases , then increases , and finally decreases again as @xmath111 approaches the dw centre ( @xmath125 ) . as we discussed already , the fluctuation of the magnetization at the dw centre is small when @xmath68 is close to @xmath115 . the spin at the boundary only interacts with one nearest spin , so it has more freedom to fluctuate . for the others , because of the influence of the dw structure ( or boundary ) , the fluctuations of the spins which are near the dw ( or near the boundary ) are larger compared to those of a spin located in the middle of a polarized region . thus @xmath93 is larger if the spin is located near the dw or near a boundary . \2 ) @xmath126 : @xmath93 reaches its maximum at the dw centre . the reason for this is that in this regime the magnetizations of all spins retain their initial direction , therefore the spins that are far from the centre fluctuate little . \3 ) @xmath127 : in this regime ( ising limit ) , the initial state is very close to the eigenstate , and the fluctuations are small , even for the spins at the dw . in the presence of ising - like anisotropy , dws in a ferromagnetic spin @xmath0 chain are dynamically stable over extended periods of time . the profiles of the magnetization of the dw are different from the profile in the ground state in the subspace of total magnetization @xmath42 . as the system becomes more isotropic , approaching the quantum critical point , the width of the dw increases as a power law , with an exponent equal to @xmath0 . c. leforestier , r.h . bisseling , c. cerjan , m.d . feit , r. friesner , a. guldberg , a. hammerich , g. jolicard , w. karrlein , h .- meyer , n. lipkin , o. roncero and r. kosloff , j. comp . phys . * 94 * , 59 ( 1991 ) .
we study the real - time domain - wall dynamics near a quantum critical point of the one - dimensional anisotropic ferromagnetic spin @xmath0 chain . by numerical simulation , we find the domain wall is dynamically stable in the heisenberg - ising model . near the quantum critical point , the width of the domain wall diverges as @xmath1 .
the satellite and ground based imaging atmospheric cerenkov telescopes such as h.e.s.s . , magic , and veritas have demonstrated that the ultra - high energy universe is teeming with energetic very high - energy gamma - ray ( vhegr , @xmath3 ) sources , the extragalactic component of which mainly consists of tev blazars with a minority population of other sources such as radio galaxies and starburst galaxies . these vhegr observations are being used to constrain the sites and mechanisms of particle acceleration ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , dynamics of black hole jets ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , and intergalactic magnetic fields ( igmf ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? while these objects have an interesting phenomenology , they are believed to have a minor impact upon the universe at large , i.e. , on the formation of structures and thermodynamics . energetically this is not an unreasonable assumption , as vhegr emission is @xmath4% of the radiative power of quasars . however , in this series of papers , we show that in spite of their energetic disadvantage , tev blazars have a significant effect on structure formation and a dominant effect on thermodynamics that is the vhegr emission from blazars `` punches '' far above its energetic weight . if _ the radiation from vhegr sources is thermalized , as we have argued is the case in ( * ? ? ? * hereafter paper i ) , the heating due to these vhegrs dominates photoionization heating throughout the vast majority of the universe at @xmath5 , raising the temperature of the low - density igm by up to two orders of magnitude . given that the total power emitted by agns and stars in the uv and x - rays vastly exceeds that due to the tev blazars , it seems counterintuitive that blazar heating dominates photoheating . however , the uv and x - ray background heats the intergalactic medium ( igm ) inefficiently after reionization , while vhegr photons heats the igm efficiently via plasma beam instabilities ( paper i ) . this difference in the heating efficiency of photoheating vs. blazar heating is due to the difference in the rate - limiting process in each . the rate of photoheating after reionization is not limited by the availability of ionizing photons , but instead by abundance of targets , and thus recombination . on the other hand , the heating of the igm by tev blazars does not suffer from a deficit of targets and is only limited by the total cosmic power of vhegr sources . we illustrate this point with the following order of magnitude estimate . first let us estimate the amount of photoheating that the igm suffers at the present day . the recombination rate of h is of order the hubble time at the mean density of the universe at present . hence , an average h atom in the igm will recombine once over a hubble time only to be ionized immediately by a uv photon . using a spectral index of -1.6 for the ionizing background , which is appropriate for quasars @xcite , the average amount of excess energy absorbed per ionization is @xmath6 . the fraction of the rest mass energy of all the baryons in the universe required to produce this amount of heating in the igm is @xmath7 hence , only a small amount of rest mass energy is injected as thermal energy into the igm . an equivalent statement is that the diffuse igm is optically thin to ionization radiation . by comparison , the fraction of the baryon rest mass locked up in massive black holes in the universe is @xmath8 and in stars is @xmath9 @xcite . assuming a radiative efficiency ( relative to rest mass ) for black holes and stars of @xmath10 and @xmath11 respectively , we find that fraction of the baryon rest mass converted to radiation in black holes and stars is @xmath12 and @xmath13 , respectively . this is many orders of magnitude larger than what is required by equation ( [ eq : excess ] ) and demonstrates an important point : _ hard ionizing radiation is inefficient at heating the igm after h and he reionization . _ in paper i , we argued that vhegr photons are efficiently converted into heat in the igm via plasma instabilities . hence , the limiting factor is the total energy density of vhegr photons over cosmic time . to estimate this energy density , we note that the local _ observed _ tev blazar luminosity density is @xmath14 that of the local quasar luminosity density ( see section [ sec : z=0 heating ] ) , after correcting for various selection effects . assuming that the tev blazar luminosity density tracks the quasar luminosity density over cosmic time , this implies that the fraction of baryon rest mass that is converted to vhegr photons ( and ultimately heating of the igm ) is @xmath15 which is nearly five times that due to photoheating ( i.e. , equation ( [ eq : excess ] ) ) and demonstrates the dominance of tev blazar heating . equivalently , the greater efficiency of tev blazar heating compared to photoheating more than makes up for its energetic disadvantage . the physics of this heating and its cosmological consequences is the subject of this series of three papers . in paper i , we studied the physics of vhegr photon propagation through the universe . as these vhegrs propagate through the universe , they interact with the soft photons that comprise the extragalactic background light ( ebl ) and produce ultra - relativistic pairs ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? typical mean free paths are between @xmath16 and @xmath17 , depending upon the energy of the vhegr and redshift , i.e. , the universe is optically thick to vhegr . the result is a ubiquitous population of ultra - relativistic pairs , with typical lorentz factors of @xmath18@xmath19 . previously , it has been assumed that they lose energy exclusively through inverse - compton scattering the cosmic microwave background ( cmb ) and ebl , producing gev gamma rays that form part of the egrb ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the non - observation of this gev gamma - rays has been used to argue for cosmologically interesting igmfs @xcite . we then presented a plausible alternative mechanism for extracting the kinetic energy of the ultra - relativistic pairs : plasma beam instabilities . despite the extraordinarily dilute nature of this ultra - relativistic pair plasma , we found a variety of plasma instabilities which grow on timescales short in comparison to the inverse - compton cooling time , the most important of which is the `` oblique '' instability @xcite . via this instability , these ultra - relativistic pairs lose their kinetic energy by depositing it as heat in the igm . because these beams then cool well before an inverse - compton cascade ( icc ) can develop , the simplest versions of the argument used to produce limits upon the igmf are precluded . hence , the existence of the igmf does not follow from the non - observation of gev gamma rays from existing tev blazars as previous groups have argued . in addition , the lack of an icc allows for a large and evolving blazar population without upsetting the limits on the egrb and statistics of high - energy blazars . based upon the plasma - instability mechanism , we now adopt as a hypothesis that the ultra - relativistic pairs primarily deposit their energy in the igm via this or a related mechanism . in this paper ( paper ii ) , we explore the impact of this heating on the thermodynamics of the igm . we estimate the amount of heating provided by the observed tev blazar population after correcting for the selection effects of the current pencil beam vhegr observations using the all - sky monitoring of the satellite . we will show that the luminosity density in vhegrs is of order @xmath20 of the quasar luminosity density , which dominates the photoheating rate at low @xmath21 . we then explore the qualitative and quantitative nature of this heating , which in essence , serves as an alternate feedback mechanism . in particular , tev blazars deposit heat evenly in a _ volumetric _ sense , i.e. , independent of the local igm density . hence , this heating deposits more energy per baryon in low - density regions than in high - density regions , naturally producing an inverted temperature - density relation in voids . with only a minor rescaling of the empirically normalized of observed blazar heating , we find that it is possible to reproduce the inferred inverted temperature - density relation at @xmath22 @xcite , something which has proven to be a problem within the context of standard reionization models @xcite . in ( * ? ? * hereafter paper iii ) , we will explore the result of this additional igm heating upon the formation of structure in the universe . in particular , we will show that the injection of entropy into the igm by tev blazars contributes to developing a redshift dependent entropy floor for galaxy clusters and groups at @xmath23 and suppresses the formation of dwarfs . we will highlight that the redshift dependent nature of tev blazar heating in our model suggests a large injection of entropy around @xmath24 , which boosts the entropy of late forming objects . this predicted enhanced entropy of young groups is consistent with recent observations that show optically bright and therefore young , groups and clusters are x - ray dim that is having a lower gas density due to a raised entropy floor . we also will show that tev blazar heating suppresses the formation of late forming dwarfs both in galactic halos , i.e. , the missing satellite problem @xcite , and in voids , i.e. , the void phenomenon @xcite by raising the temperature of the igm such that gas can not collapse to form galaxies . this work is organized as follows : we first review the fate of energy carried by vhegr photons in section [ sec : pair propagation ] , discussed in detail in paper i. we describe how vhegr photons produce pairs in the igm and how these ultra - relativistic pair beams are unstable to plasma instabilities . in particular , we highlight the `` oblique '' instability , which is especially efficient at converting the kinetic energy of the beams into thermal energy in the igm . motivated by our review of paper i , we adopt the assumption that the kinetic energy of the ultra - relativistic pairs is thermalized in the igm , either via the `` oblique '' instability or some related mechanism . in section [ sec : igmhr ] , we estimate the current tev - blazar igm heating rate by collating the known extragalactic tev blazars with a well measured spectrum , accounting for incompleteness ( section [ sec : z=0 heating ] ) . we use the similarity between the luminosity functions of nearby quasars and tev blazars found in paper i , to extend the heating rate to @xmath25 and estimate the tev - blazar covering fraction ( sections [ sec : z>0 heating ] and [ sec : homogeneity ] ) . the implications for the thermal history of low - density regions ( less than 10 times the mean density , i.e. , @xmath26 ) are explored in section [ sec : thermal history ] . generally , we find that without any fine tuning it is possible to reproduce the inverted temperature - density relation at @xmath22 inferred by high - redshift studies @xcite , while simultaneously satisfying the temperature constraints at @xmath27 ( e.g. , those by * ? ? ? * ) and leaving the local forest unaffected . the results of this work and paper iii assume that the energy of tev blazars are efficiently thermalized in the igm . though we have identified a particularly promising instability , i.e. , the `` oblique '' instability , in paper i , the particular details by which the energy in ultra - relativistic pairs is thermalized are unimportant . instead , the results of this work and paper iii depend solely on the gross energetics of the tev blazars . as a consequence , observations of the thermal history and constraints upon the structures in low - density regions represent an independent empirical probe of the fate of the ultra - relativistic pairs produced by tev blazars . for all of the calculations presented below ( and in this series ) we have assumed the wmap7 cosmology with @xmath28 , @xmath29 , @xmath30 , and @xmath31 @xcite . the observed extragalactic vhegr sources are all located at low redshift ( @xmath32 ) . this is a due to the annihilation of vhegrs on the ebl , producing ultra - relativistic @xmath33 pairs ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? here we describe the fate of these vhegrs , and the consequences of propagating through the intergalactic medium ( igm ) for the pairs they produce . we refer the interested reader to sections 2 and 3 of paper i for a more complete discussion of the properties of the generated pairs and the importance and nature of plasma beam instabilities upon their propagation . vhegrs are attenuated by the pair production off ebl photons . namely , when the energies of the vhegrs ( @xmath34 ) and the ebl photon ( @xmath35 ) exceed the rest mass energy of the @xmath33 pair in the center of mass frame , i.e. , @xmath36 , where @xmath37 is the relative angle of propagation in the lab frame , an @xmath33 pair can be produced with lorentz factor @xmath38 @xcite . an estimate for the mean free path of vhegr photons is ( paper i ) @xmath39 where the redshift evolution is due to that of the ebl alone , is dependent predominately upon the star formation history , and @xmath40 for @xmath41 and @xmath42 for @xmath43 @xcite and declines rapidly afterward , galaxies and type 1 agns compensate for the lost flux until @xmath44 . see , for example , figure 3 from @xcite . ] . relative to the hubble length , the attenuation length of vhegrs is very short . in fact , above @xmath45 the universe is optically thick to sources at @xmath46 ( cf . * ? ? ? * ) . since @xmath47 is much larger than any conceivable source size , and @xmath48 , locally these pairs necessarily constitute a cold , highly anisotropic beam . the production of pairs by the interaction of vhegrs with the ebl is opposed by the removal of these pairs by various cooling processes . that is , the evolution of the characteristic pair - beam density at the injection lorentz factor , @xmath49 , is governed by the boltzmann equation : @xmath50 where the left - hand side assumes all the pairs are moving away from the vhegr source relativistically ( @xmath51 and @xmath52 ) , the right - hand side corresponds to pair production , and @xmath53 is the cooling / removal rate of these pairs . in a homogeneous steady state , the rate of production is balanced by the rate of removal , which gives : @xmath54 where the rate of production is given by @xmath55 . generally , @xmath53 is a function of energy and beam energy and refer the reader to paper i for additional details . however , for any choice of @xmath53 , the solution to equation ( [ eq : nb ] ) gives @xmath56 . commonly , it is assumed that the pairs evolve primarily due to an icc which deposits the energy from the original vhegr photon near @xmath57 . note that this is fundamentally radiative ; after the vhegrs are scattered down to @xmath45 they effectively decouple from the universe . iccs of these pair beams widely exploited as a possible probe of igm magnetic fields in the context of the missing inverse compton features at @xmath45 ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the associated cooling rate for this process for pairs with lorentz factor @xmath58 is @xmath59 where @xmath60 denotes the thompson cross section . the strong redshift dependence arises from the rapid increase in the cmb energy density with @xmath21 ( @xmath61 ) . however , this may not be the dominant process . as we showed in paper i , plasma beam instabilities are a potential mechanism by which the kinetic energy of the pairs is extracted much more rapidly . the two plasma beam instabilities most frequently discussed , the two stream and weibel instabilities , are particular limits of a more general instability , differentiated by the direction of the perturbed wave vector with respect to the beam orientation ( perpendicular for weibel , parallel for the two stream ; * ? ? ? . for dilute beams , by far the most powerful growth occurs at oblique angles , and thus named the `` oblique '' instability . paper i and @xcite gives a more extensive discussion of these three plasma instabilities . relative to the laboratory frame , these pair beams are `` cold '' , i.e. , their transverse momentum is much smaller than their parallel momentum , but even the small transverse temperatures that are acquired from pair production for these pair beams are important , placing the oblique instability in the kinetic regime . here , the oblique instability cooling rate has been numerically measured to be @xmath62 where @xmath63 is the igm free - electron number density ( assuming full ionization ) , @xmath64 is the elementary charge , and we have set the beam temperature in the beam frame to @xmath65 , characteristic of that induced by pair production , and thus @xmath66 @xcite . both the cold and hot growth rates have been verified explicitly using particle - in - cell ( pic ) simulations , though at somewhat less dilute beams than the pair beams from tev blazars @xcite . for the kinetic oblique instability , the cooling rate is @xmath67 where @xmath68 is the isotropic - equivalent luminosity per unit energy of the vhegr source . this is a stronger function of photon energy than inverse - compton cooling , implying that it will eventually dominate at sufficiently high energies , assuming a flat vhegr spectrum . in addition it is a very weak function of @xmath69 , being only marginally faster in lower - density regions , and thus the cooling of the pairs is largely independent of the properties of the background igm . the effective cooling rates induced by the `` oblique '' instability can be obtained by numerically solving equation ( [ eq : nb ] ) for @xmath49 , with @xmath70 as described in paper i. for the purpose of this paper , for tev blazars the effective cooling rate is generally dominated by the `` oblique '' mode and thus the fraction of vhegr emission that is effectively thermalized is nearly unity . therefore , we are justified in assuming that all of the vhegr emission is thermalized . in section [ sec : z>0 heating ] we will briefly revisit this assumption at high @xmath21 , finding again that it is well justified at all @xmath21 . however , this may not be the case for all potential vhegr sources . the physics of the `` oblique '' instability , like all plasma beam instabilities , depends strongly upon the beam density . for the pair beams that result from vhegr - ebl photon interactions , the beam density is a function of source luminosity . hence for sufficiently low luminosity systems , the `` oblique '' mode grows so slowly that inverse compton off of the cmb dominates it . this defines a critical isotropic - equivalent luminosity for plasma beam instabilities to be relevant , typically near @xmath71 , though depending upon redshift ( see figure 3 of paper i ) . for the tev blazars considered in this paper , this is generally not a concern . however , this will be important in our discussion of alternative vhegr sources in section [ sec : alternatives ] . as we have argued in paper i and reviewed in section [ sec : pair propagation ] , plasma instabilities on the pair beam dissipate the bulk of the vhegr emission . in particular , the `` oblique '' instability appears to be a promising mechanism by which the ultra - relativistic pair beams that are produced from the interaction of vhegr and ebl photons blazars are efficiently thermalized and converted to local igm heating . henceforth , we will make the assumption for the rest of this work and again in paper iii that the energy of tev blazars is efficiently thermalized in the igm . with this stated assumption , we now discuss the sources that dominate the igm heating rate , estimate the magnitude of the heating rate and its evolution with redshift , and briefly assess the homogeneity with which this new process occurs . the local heating rate associated with the dissipation of the high - energy emission from a single blazar is determined by two factors : the rate at which the vhegrs are converted into pairs and the rate at which the energy of the pairs is subsequently converted locally into heat . the former is determined by the pair - production cross - section . due to the efficiency of the plasma instabilities in dissipating the beam energy , the latter is set by the ratio of the relevant cooling rates . that is , the single - blazar heating rate is @xmath72 where @xmath73 is a dimensionless function due to the pair - creation threshold and which depends upon the shape of the ebl spectrum ( we set @xmath73 to vanish for @xmath74 and be unity otherwise ) , and @xmath75 which is a function of @xmath76 , @xmath34 , and @xmath21 via the dependence of the cooling rates upon @xmath49 , @xmath58 , and @xmath21 . generically , @xmath77 represents the fraction of pair energy that is thermalized , but we have chosen a specific form of @xmath77 corresponding to the `` oblique '' instability to make the discussion below more concrete . while it is also very weakly dependent upon @xmath69 , this may be neglected in practice . within the linear regime the plasma instabilities responsible for the dissipation of pair beam energy are independent of those arising from beams in substantially different directions . since the local vhegr flux is dominated by a number of sources that is small in comparison to that needed to isotropize the beam s phase space , we may treat the resulting evolutions of their associated pair beams independently and sum their resulting heating rates to determine the total heating rate , @xmath78 . ( see paper i ) ; hence we would need @xmath79 blazars to isotropize any given point in space . however , as will be shown in section [ sec : homogeneity ] the number of blazars that contribute substantially to the local heating rate is much smaller ; hence we believe the local pair distribution function is sufficiently anisotropic . ] the local heating rate can be estimated in at least two ways , both of which give similar results . first , since the high - energy gamma rays deposit their energy locally , we can identify the local heating rate with the high - energy gamma - ray luminosity density of tev blazars . in paper i we showed that this is roughly @xmath80 that of quasars , and thus approximately @xmath81@xmath82@xmath83 , depending upon the minimum luminosity at which the heating mechanism operates . alternatively , given a sufficiently complete sample , we can estimate the local heating rate using that implied by the fluxes of the observed tev blazars . in the present epoch , @xmath84 at the relevant @xmath34 , and thus the heating rate takes the particularly simple form : @xmath85 to evaluate this sum , we have collated the presently 46 extragalactic vhegr sources known.rwagner/sources/ for an up - to - date list . ] of these 46 , only 28 have published measurements of their vhegr flux from a combination of veritas , h.e.s.s . , and magic observations . for these 28 sources , we have extracted the parametrized spectra assuming the form , @xmath86 where @xmath87 is the normalization in units of @xmath88 . the gamma - ray energy flux is trivially related to @xmath89 by @xmath90 , from which we obtain a vhegr flux , @xmath91 and for sources with a measured redshift a corresponding isotropic - equivalent luminosity , @xmath92 , where @xmath93 is the luminosity distance . the resulting @xmath87 , @xmath94 , @xmath95 , @xmath96 , and @xmath97 are collected in table [ tab : tevsources ] . in addition we list the redshift , inferred distance , and absorption - corrected intrinsic spectral index at @xmath94 , obtained via @xmath98 where @xmath99 is the optical depth accrued by a vhegr emitted at redshift @xmath21 and with observed energy @xmath34 . in paper i , where there we set @xmath34 to the _ emitted _ energy of the gamma ray . a full definition at arbitrary observer redshift can be found in equation ( [ eq : tau ] ) . ] for high - redshift sources @xmath100 can be less than @xmath101 , implying that an intrinsic spectral upper - cutoff must exist . here we conservatively take this to be at @xmath102 , which is well justified given the distances to the two sources that dominate the observed local @xmath103 flux , mkn 421 and 1es 1959 + 650 , though as we shall see below , the heating rate is relatively insensitive to the particular values of the lower and upper spectral cutoffs . lccccccccccl mkn 421 & 0.030 & 129 & 68 & 1 & 3.32 & @xmath104 & 45.6 & 3.15 & 44 & h & @xcite + 1es 1959 + 650 & 0.047 & 201 & 78 & 1 & 3.18 & @xmath105 & 45.9 & 2.90 & 47 & h & @xcite + 1es 2344 + 514 & 0.044 & 190 & 120 & 0.5 & 2.95 & @xmath106 & 45.0 & 2.82 & 9.2 & h & @xcite + mkn 501 & 0.034 & 150 & 8.7 & 1 & 2.58 & 85 & 44.4 & 2.39 & 6.0 & h & @xcite + 3c 279 & 0.536 & 2000 & 520 & 0.2 & 4.11 & 68 & 46.9 & 2.53 & 1.0 & q & @xcite + pks 2155 - 304 & 0.116 & 490 & 1.81 & 1 & 3.53 & 64 & 45.4 & 2.75 & 1.4 & h & @xcite + pg 1553 + 113 & @xmath107 & @xmath108 & 46.8 & 0.3 & 4.46 & 41 & @xmath109 & @xmath110 & @xmath111 & h & @xcite + w comae & 0.102 & 430 & 20 & 0.4 & 3.68 & 31 & 44.9 & 3.41 & 0.6 & i & @xcite + 3c 66a & 0.444 & 1700 & 40 & 0.3 & 4.1 & 28 & 46.3 & 2.43 & 0.4 & i & @xcite + 1es 1011 + 496 & 0.212 & 870 & 200 & 0.2 & 4 & 26 & 45.5 & 3.66 & 0.4 & h & @xcite + 1es 1218 + 304 & 0.182 & 750 & 11.5 & 0.5 & 3.07 & 24 & 45.4 & 2.37 & 0.8 & h & @xcite + mkn 180 & 0.045 & 190 & 45 & 0.3 & 3.25 & 20 & 44.0 & 3.17 & 0.6 & h & @xcite + 1h 1426 + 428 & 0.129 & 540 & 2 & 1 & 2.6 & 20 & 45.0 & 1.71 & 1.4 & h & @xcite + rgb j0710 + 591 & 0.125 & 520 & 1.36 & 1 & 2.69 & 15 & 44.8 & 1.83 & 0.9 & h & @xcite + 1es 0806 + 524 & 0.138 & 580 & 6.8 & 0.4 & 3.6 & 10 & 44.7 & 3.21 & 0.2 & h & @xcite + rgb j0152 + 017 & 0.080 & 340 & 0.57 & 1 & 2.95 & 8.5 & 44.1 & 2.45 & 0.3 & h & @xcite + 1es 1101 - 232 & 0.186 & 770 & 0.56 & 1 & 2.94 & 8.2 & 44.9 & 1.50 & 0.3 & h & @xcite + 1es 0347 - 121 & 0.185 & 770 & 0.45 & 1 & 3.1 & 8.2 & 44.9 & 1.67 & 0.3 & h & @xcite + ic 310 & 0.019 & 83 & 1.1 & 1 & 2.0 & 8.1 & 42.8 & 1.90 & 0.1 & h & @xcite + pks 2005 - 489 & 0.071 & 300 & 0.1 & 1 & 4.0 & 8.0 & 44.0 & 3.56 & 0.1 & h & @xcite + magic j0223 + 430 & & & 17.4 & 0.3 & 3.1 & 7.6 & & @xmath112 & 0.2 & r & @xcite + 1es 0229 + 200 & 0.140 & 590 & 0.7 & 1 & 2.5 & 6.4 & 44.5 & 1.51 & 0.5 & h & @xcite + pks 1424 + 240 & @xmath113 & @xmath114 & 51 & 0.2 & 3.8 & 6.3 & @xmath115 & @xmath116 & 0.1 & i & @xcite + m87 & 0.0044 & 19 & 0.74 & 1 & 2.31 & 5.9 & 41.4 & 2.29 & 0.6 & r & @xcite + bl lacertae & 0.069 & 290 & 0.3 & 1 & 3.09 & 5.4 & 43.8 & 2.67 & 0.2 & l & @xcite + h 2356 - 309 & 0.165 & 690 & 0.3 & 1 & 3.09 & 5.4 & 44.6 & 1.86 & 0.2 & h & @xcite + pks 0548 - 322 & 0.069 & 290 & 0.3 & 1 & 2.86 & 4.0 & 43.7 & 2.44 & 0.2 & h & @xcite + centaurus a & 0.0028 & 12 & 0.245 & 1 & 2.73 & 2.8 & 40.7 & 2.72 & 0.2 & r & @xcite + ) and lower - cutoff energies ( red , black , and blue lines ) . in particular , note that the overall value of @xmath117 is relatively insensitive to reasonable variations in the spectral cutoffs . ] using the observed vhegr source in table [ tab : tevsources ] , the heating rate associated with the 23 ibl / hbl blazars ( labeled i and h in the table , see below ) is @xmath118 note that for @xmath119 , characteristic of the two dominant @xmath103 sources , this depends logarithmically upon the lower and upper spectral cutoffs . the resulting @xmath117 is shown in figure [ fig : qobs ] as a function of the upper spectral cutoff , @xmath120 for a handful of different lower cutoffs , ranging from @xmath121 to @xmath122 . for our fiducial values , @xmath123 , with this only weakly depending upon the various cutoffs , changing by at most by a factor of 2 over the reasonable ranges . hence , for the remainder of this work , we choose a lower spectral cutoff of 100 gev . to complete this estimate , we now correct for the selection effects of vhegr observations . to correct for the pointed nature of vhegr observations , we rely on the all - sky @xmath124 gamma - ray observations from the fermi satellite @xcite of tev blazar . those belong to 2 subclasses of blazars , namely high - energy peaked bl lacs ( hbl ) and the somewhat less efficient accelerators , intermediate - energy peaked bl lacs ( ibl ) which are , in some cases , also able to reach energies beyond 100 gev . outside of the galactic plane , observes 118 high - synchrotron peaked ( hsp ) blazars and a total of 46 high - synchrotron peaked ( isp ) blazars . roughly half of the latter are likely to emit vhegrs as indicated by their flat spectral index between 0.1 and 100 gev , @xmath125 ( see the spectral index distribution of figure 14 in @xcite ) . of these potential 141 tev blazars , only 22 have also been coincidentally identified as tev blazars ( out of a total of 28 coincident tev sources ) , while there are a total of 33 known tev blazars ( 29 hbl , 4 ibl ) . if these 141 sources are all @xmath103 emitters , but have not been detected due to incomplete sky coverage of current tev instruments , then the selection factor is @xmath126 . in addition , the duty cycle of coincident @xmath124 and @xmath103 emission is @xmath127 . here we assume that the luminosity distribution of observed vhegr sources reflects the true distribution after correcting for the effects of flux limitations in the observations . that this may be done is justified empirically in section 5.1.2 of paper i , and demonstrated explicitly in figure 5 of paper i. thus , we may use constant correction factors ( independent of luminosity ) to estimate the true distribution . finally , by excluding the galactic plane for galactic latitudes @xmath128 , this is an underestimate by roughly @xmath129 . taken together , the true heating rate in our `` standard '' model is then : @xmath130 where @xmath131 is a remaining coefficient of order unity correcting for systematic uncertainties . these include corrections to the spectral model we currently use and the corrections to the completeness of the fermi sample of hbls and ibls in accounting for the complete population of vhegr sources . the fact that there are already 4 radio galaxies of faranoff - riley type i ( fr i ) , 2 flat spectrum radio quasars , and 2 starburst galaxies emitting vhegrs implies that we are probably too conservative in accounting for all the vhegr sources . to explore the effects of the variation in the heating rate , we also adopt an `` optimistic '' heating model that is normalized to fit the observed igm inverted temperature - density relation @xcite . this `` optimistic '' model has @xmath132 yielding a heating rate of @xmath133 . in estimating @xmath78 from equation ( [ eq : qdotest2 ] ) we have made a number of implicit assumptions . first , we have assumed that the observed distribution of tev blazars locally is representative of the average distribution at _ any given point _ in the universe . since the blazars that have been observed locally go out to @xmath134 , we believe this is a relatively safe assumption . second , we have assumed that the current sample of blazars is sufficiently flux - complete to dominate the heating rate at earth . that is , we have assumed that the effective flux limit of the current generation of cerenkov telescopes is low enough that it captures the bulk of the sources responsible for the local heating of the igm . that this is the case is less clear , and will be the subject of section [ sec : homogeneity ] and the appendix in some detail . unfortunately , since the constant of proportionality relating the local tev blazar and quasar luminosity densities is obtained using the observed set of blazars , our two estimates are strongly correlated . nevertheless , this provides a convenient lower limit upon @xmath78 , and we use this heating rate , which we denote the `` standard '' heating rate , as the mean heating rate of any fluid element in the present - day universe . over a hubble time , the total heat deposited per unit volume is then roughly @xmath135 which is sufficient to raise the temperature of @xmath136 regions to approximately @xmath137 , and thus dramatically alter the thermal history of voids . extrapolating the above estimate to @xmath25 requires understanding how the tev blazar properties and population have evolved , as well as the evolution in @xmath138 . generally , the average heating rate is given by @xmath139 where @xmath140 is the physical number density of blazars at a given redshift per unit logarithmic isotropic - equivalent luminosity , spectral index , and blazar jet opening angle ( we ve assumed all jets are symmetric ) . we make the simplifying assumption that @xmath141 is separable into components describing the evolving luminosity density distribution , @xmath142 , and a static , unit - normalized spectral distribution , @xmath143 , where @xmath144 . this produces , @xmath145 where the bulk of the redshift and luminosity dependence is now contained in the outer - most integral over @xmath146 , and we have used @xmath147 ( since the heating is dominated by nearby objects , for simplicity we do not distinguish between different cosmological distance definitions , though see the appendix for a more careful treatment ) . if there were no redshift or flux dependence in the remaining terms , it would be possible to simply normalize the heating rate by @xmath148 and the estimated tev - blazar luminosity density , which we define to be @xmath149 where @xmath150 is chosen so that the plasma instabilities operate efficiently at all redshifts of interest . however , this is not entirely the case since @xmath138 retains some dependence upon @xmath21 , and thus we set @xmath151 where @xmath152 provides a correction due to the changing strength of the pair beam dissipation mechanism in comparison to inverse - compton cooling . ( black solid ) , @xmath153 ( blue long - dashed ) , and @xmath154 ( red short - dashed ) . for comparison , the tev blazars that dominate the heating locally have luminosities of roughly @xmath155 . ] the vhegr flux at earth is dominated by a handful of very bright sources for which the absorption corrected @xmath119 at @xmath156 and isotropic - equivalent luminosity @xmath157 . this simplifies the estimation of @xmath152 significantly , giving @xmath158 in which the form and magnitude of @xmath68 is fixed . the resulting correction factors are shown in figure [ fig : qcorr ] for isotropic - equivalent luminosities above @xmath45 ranging from @xmath154 to @xmath159 . generally , @xmath152 makes a small ( @xmath160 ) correction following the peak of the quasar luminosity function ( @xmath27 ) , though can grow to as much as @xmath161 by @xmath162 for dim objects . however , since we will be primarily interested in bright blazars at @xmath163 , @xmath152 modifies the heating rate at relevant redshifts by @xmath164 ( see also section [ sec : tbprop ] ) . thus , given the comparatively larger uncertainties associated with the tev blazar luminosity function , plasma heating mechanism , and additional potential vhegr sources , we neglect @xmath152 in what follows . ) as a function of redshift . the shaded region represents the 1-@xmath165 uncertainty that results from a combination of the uncertainty in the number of bright blazars that contribute to the local heating and in the uncertainties in the @xcite quasar luminosity density to which we normalize . our optimistic and standard models , defined in section [ sec : z=0 heating ] , are shown by the magenta short - dash - dot and blue long - dash - dot lines . insert : comparison between the observed blazar luminosity function ( data points with one sigma error bars and upper limits ) with the quasar luminosity function that has been shifted by a factor of @xmath14 in luminosity density and @xmath166 in luminosity . this insert is a simplified reproduction of figure 5 of paper i. [ f : bld ] ] therefore , obtaining the heating rate as a function of @xmath21 is reduced to estimating @xmath167 , the normalized evolution of the blazar luminosity density . however , no vhegr emitting blazars are known beyond @xmath168 , presumably due to the pair - creation absorption associated with propagation through the ebl . as a consequence , nothing is known about @xmath169 at high redshifts directly . nevertheless , in paper i we showed that a close relationship between @xmath170 and the quasar luminosity function of @xcite , @xmath171 , exists at low @xmath21 : @xmath172 the insert in figure [ f : bld ] shows a comparison between the observed tev - blazar luminosity function , calculated from the distribution of the sources in table [ tab : tevsources ] and applying an empirically determined flux limit of @xmath173 , with @xmath174 ( for a full description of how the luminosity function was formed , and how it compares to @xmath175 , see section 5 of paper i ) . furthermore , we showed that once the iccs are suppressed by the plasma beam instabilities ( or some analogous mechanism ) , extending this relationship to high-@xmath21 was in excellent agreement with the best current constraints upon the high-@xmath21 tev blazar population : the tev blazar statistics and the measurement of the extragalactic gamma - ray background ( between @xmath176 and @xmath45 ) . thus , we estimate @xmath177 , shown in figure [ f : bld ] , where @xmath178 is the luminosity density of quasars . figure [ f : bld ] shows that employing the quasar luminosity density to estimate the heating rates has profound consequences . in particular , the inferred luminosity density of tev blazars ( solid line ) rises rapidly with increasing @xmath21 , with the comoving density increasing by roughly a factor of @xmath179 by @xmath27 ( see also figure 8 of * ? ? ? * ) . in physical units this corresponds to a increase by a factor of nearly @xmath180 . thus , we expect an increase in the local heating rate by a similar factor over the presently estimated value near @xmath27 . + we now investigate the assumption of even heating as a prelude to our simple one - zone model of the igm . in lieu of large - scale simulations , the homogeneity of the heating due to tev blazars is difficult to assess for a variety of reasons . first , the duty cycle of the tev blazars is unknown . second , the density at high redshifts is poorly constrained . third , the importance of clustering bias is unclear . fourth , it is difficult even to define which blazars are relevant , e.g. , which luminosity range contributes the bulk of the local heating . nevertheless , we make an attempt to roughly characterize the possible range in the stochasticity of the local heating rates via a number of different estimates . , in the @xmath45@xmath181 band ) above @xmath182 ( dot - dash ) , @xmath154 ( solid ) , @xmath153 ( short - dash ) , @xmath159 ( long - dash ) , and the local luminosity - weighted median luminosity of tev blazars ( red dot ) , as functions of redshift . for reference the spectrally averaged ( with @xmath183 ) mean free path , @xmath184 , and the local mean free path for a @xmath156 gamma ray are shown by the grey solid and dashed lines , respectively . for sources with a spectral break between @xmath45 and @xmath156 , we anticipate the effective mean free path to lie between these . ] we begin by estimating the instantaneous comoving mean separation of _ visible _ blazars above various luminosity thresholds , @xmath185^{-1/3}\,,\ ] ] where the precise value of @xmath186 is unimportant as long as it is above the peak of the luminosity function ; here we choose @xmath187 , consistent with the theoretically expected upper limit of tev blazar luminosities . since we have determined @xmath170 from the observed blazar population , and are making use of the isotropic - equivalent luminosities , this is independent of the blazar jet opening angle ; smaller opening angles will result in a correspondingly larger number of objects such that the total number seen is unchanged , and therefore @xmath188 is fixed . a comparison between @xmath188 and the mean free path of vhegrs is shown in figure [ f : bms ] . because @xmath47 varies dramatically from @xmath45 to @xmath181 , for this purpose , we define a spectrally averaged mean free path , @xmath184 , determined implicitly by @xmath189 in which we choose @xmath183 based upon the local tev blazar sample . this is roughly the @xmath64-folding distance of the entire spectral band , shown in figure [ f : bms ] by the grey solid line , and in practice is quite close to @xmath190 . if the intrinsic tev blazar spectra peak above @xmath45 , our estimate of the typical vhegr mean free path could be significantly too large ; thus we also compare @xmath188 to @xmath47 at @xmath156 ( the grey dashed line in figure [ f : bms ] ) , providing an extreme lower - bound upon the spectrally averaged mean free path in practice . the rapid increase of the density of ebl photons with @xmath21 , associated with the larger star formation rate in the recent past , results in a substantially reduced @xmath47 by @xmath44 . prior to @xmath44 , the physical number density of ebl photons remains nearly constant , and thus @xmath191 in comoving units . in the present epoch , @xmath184 is quite large , and thus despite their sparsity the mean separation of even bright blazars ( @xmath159 ) is is less than @xmath184 . this remains true for objects with luminosities @xmath192 , which includes the local luminosity - weighted median tev blazar luminosity , @xmath193 , defined such that @xmath194 hence , locally , we expect blazar heating to be quite uniform . at high redshift matters change somewhat . until @xmath44 , @xmath195 and @xmath196 both decrease more rapidly than @xmath188 . for @xmath46 , in comoving units @xmath195 and @xmath196 increase slowly , though at a marginally larger rate than @xmath188 . as a result , near @xmath197 the mean separation of tev blazars is largest in comparison to the vhegr mean free path . thus , the mean separation between objects more luminous than @xmath159 is larger than @xmath184 near @xmath197 . nonetheless , when the lower luminosity limit is dropped to @xmath153 or below , generally @xmath198 for all @xmath199 . similarly , the mean separation between blazars more luminous than @xmath193 is also smaller than our estimate of @xmath184 for the relevant redshift range . hence we may expect that nearly all patches of the universe will be illuminated by at least one luminous tev blazar following @xmath200 . include : the number of blazars with intrinsic isotropic - equivalent luminosities above @xmath153 within the @xmath201 surface ( blue short - dash ) , the number of blazars with luminosities above that luminosity - weighted median value at each redshift within the @xmath201 surface ( red dot ) , and the number of blazars with individual heating rates that exceed that above which 50% and 75% of the total heating is produced ( black solid and long - dash , respectively ) . for comparison , our estimate of the completeness - corrected number of tev blazars that are presently observable _ in the tev _ is shown by the filled back point , with error bars denoting the poisson uncertainty only . ] closely related to the mean separation is the number of blazars within the @xmath202 surface ( defined explicitly in the appendix ) above some luminosity limit . figure [ f : numblazars ] shows these for objects with luminosities above @xmath153 ( blue short - dash ) , and @xmath203 ( red dot ) . while these may be roughly inferred from the associated mean separations in figure [ f : bms ] , they differ slightly from @xmath204 due to the rapidly evolving @xmath184 and blazar population combined with the finite look - back time in the integral . at @xmath205 both are well above unity , with the luminosity - weighted median value substantially exceeding our estimate of roughly 170 visible extragalactic tev blazars . we have listed 23 tev blazars , but a correction factor of 7.5 needs to be applied to account for the incompleteness of present tev surveys . ] however , since generally @xmath206 not all of the blazars that contribute to the local heating are expected to appear as strong tev sources . the low-@xmath21 behavior of the number of blazars is dictated by the rapid evolution in @xmath47 , associated with the recent rapid variation in the star formation rate ( and thus the number density of ebl photons ) . prior to @xmath44 , the ebl density , and hence @xmath47 , is roughly constant in physical units , and the evolution number of visible blazars becomes indicative of the underlying evolution of blazar population . at all @xmath199 these estimates of the number of blazars responsible for the bulk of the heating , @xmath207 , exceed unity , implying only a small fractional spatial variation in the blazar heating rate . however , while the luminosity - weighted median estimate of @xmath207 does give some idea of which population is responsible for most of the tev blazar luminosity density , neither encapsulates the population responsible for the majority of the local heating . thus , if the local heating were dominated by a handful of luminous sources , it is possible for the stochasticity to be much larger . therefore , we show a third estimate of @xmath207 , corresponding to the number of sources with @xmath208s larger than the heating - rate - weighted median value ( see the appendix for a precise definition ) . that is , at a given redshift , the @xmath207 sources with the highest @xmath208s generate half of the total heating rate . thus , if a single source were to dominate the local heating rate , @xmath207 would be considerably less than unity , indicating a large degree of variability . however , below @xmath209 this is not the case ; the fractional heating rate - defined @xmath207 is considerably larger than unity . at high redshifts the heating rate becomes increasingly dominated by more distant , luminous , and rarer objects , departing from the previous estimate of @xmath207 . nevertheless , while for @xmath210 few sources contribute nearly half of the heating rate , the total heating rate must be comparable to the total vhegr luminosity density of the blazars . therefore , the total number of contributing objects must be similar to the number of visible blazars above the luminosity - weighted median @xmath97 ( i.e. , the red dotted line in figure [ f : numblazars ] ) . as a consequence , the number of relevant sources must rapidly rise with decreasing fractional heating rate . this is explicitly indicated by the dashed line in figure [ f : numblazars ] , which shows the number of sources responsible for 75% of the local heating rate . in any case , it is clear that for the redshifts at which blazar heating is likely to be important , @xmath211 , the heating rate will be relatively uniform . between @xmath212 it may experience order 50% fluctuations , and by @xmath200 will exhibit order unity deviation . however , we note that all of our estimates of the heating inhomogeneity are predicated upon the assumption that the @xmath170 accurately represents the distribution of blazars , and has the associated considerable uncertainties . for a more thorough discussion of different estimates of the number of contributing blazars , and a discussion of which blazars dominate the heating rate at a given redshift , see the appendix . in addition to the statistics of blazars , the homogeneity also depends upon the intrinsic variability of tev blazars . during the time that the tev blazars have been observed their vhegr emission has remained remarkably stable . however , this provides only a weak lower limit ( @xmath213 ) upon the variability timescale in these objects @xcite . in addition to depending upon the properties of the source itself , the variability of tev blazars depends upon the primary emission mechanism responsible for the vhegr component of tev blazars . there are two classes of inverse - compton models to explain the vhegr emission : 1 . _ synchrotron self - compton ( ssc ) model : _ in this model , the synchrotron radiation field is compton up - scattered to tev energies by a relativistic electron population @xcite . recent work has led to the conclusion that a simple homogeneous , one - zone , ssc model can not explain the sed of the majority of blazars ( see figure 36 of * ? ? ? however , models with multiple ssc components are consistent with the blazar seds . typically these invoke a steady , primary ssc component which peaks at the ir / optical ( s ) and @xmath58-ray band ( ic ) , and a second more energetic and usually more variable component , which peaks in the uv or x - ray band ( s ) and at gev / tev energies ( ic ) . the variability of this energetic component increases the probability that a given patch of the igm will see a tev blazar during its history , resulting in larger homogeneity in the blazar heating . external radiation compton ( erc ) scenario : _ this model proposes that the relativistic jet electrons compton scatter an external radiation field @xcite from the accretion disk or the dusty torus surrounding it . in the first case , the disk generates uv seed photons which are then reflected toward the jet by the broad line region within a typical distance from the accretion disk of the order of 1 pc . in the second case , the dusty torus could provide ir seed photons that are emitted at larger distances from the jet . in any of these erc models , the vhegr emission is expected to be very steady , implying long - lived tev blazars , resulting in more patchy blazar heating . we leave detailed studies of the inhomogeneity in the blazar heating rate resulting from intrinsic inhomogeneity and variability in the spatial distributions of the blazars themselves for future work . + as defined in section [ sec : rtbs ] for the tev blazars responsible for half of the total heating rate for a number of redshifts ( top : from violet blue to light - blue : @xmath214 ( solid ) , @xmath10 ( long - dash ) , @xmath215 ( short - dash ) , @xmath216 ( long - dash - dot ) , and @xmath217 ( short - dash - dot ) ; bottom : from green to dark - red : @xmath218 ( solid ) , @xmath219 ( long - dash ) , @xmath220 ( short - dash ) , @xmath221 ( long - dash - dot ) , and @xmath222 ( short - dash - dot ) ) . for reference , @xmath223 is shown in the inset ( in comoving units ) . ] in our discussion of the homogeneity of blazar heating we have necessarily attempted to define a class of objects responsible for the bulk of the heating . this is most directly done by considering those tev sources which produce , say , 50% of the total heating rate at a given observer redshift . given this population , we may also now address the properties of the most relevant sources themselves ; i.e. , we can identify which types of tev blazars dominate the heating . as we have already mentioned , this can not be too different from the set of blazars which dominate the luminosity density . we address this question in some detail in the appendix , including constructing a simple analytical toy model . figure [ f : dqdl ] shows the luminosity distribution of the heating - rate integrand once integrated over distance and assuming our form for @xmath224 ( see the appendix for a precise definition ) . the upper - half of the heating rate is dominated by sources with luminosities of approximately @xmath153 at low redshifts , and rises to @xmath159 by @xmath225 before declining slightly thereafter . this evolution is driven both by the changing shape and the changing normalization of @xmath170 with redshift . at all redshifts this is larger than the median luminosity . thus , generally it appears that the heating is due predominantly to high - luminosity objects . due to the relatively small @xmath184 , these are also necessarily nearby . up to now , our focus on the tev - blazars is motivated by the fact that the vast majority of extragalactic vhegr sources observed are blazars . other potential source of vhegr emission exist and so we will discuss these sources in this section . in assessing their potential contribution in vhegrs , however , we find that these vhegr sources are subdominant . starburst galaxies are characterized by the presence of rapid star formation , and hence , are sites of numerous supernovae . supernovae are known to be efficient particle accelerators , and are presumed to be the primary source of the galactic cosmic rays . as cosmic rays can produce vhegrs , starburst galaxies are potentially bright vhegr sources . while many starburst galaxies contain active agn , that may also emit vhegrs , unambiguous vhegr emission form the starburst itself has been seen in at least two cases , m82 and ngc 253 , with @xmath0 luminosities of @xmath226 and @xmath227 , respectively @xcite . the luminosity of these two sources are dwarfed by the local tev blazars , and thus the vhegr flux from m82 and ngc 253 does not contribute to the heating of the local igm . in addition , their vhegr luminosity produces a pair beam which is too dilute for plasma processes such as the `` oblique '' instability to beat cooling via inverse compton . hence , they do not contribute to the local vhegr flux . however , very high star - forming systems such as the ultra - luminous infrared galaxies ( ulirgs ) where @xmath228 ) and @xmath229 is the bolometric infrared luminosity ( @xmath230 ) , may evade this constraint . if the vhegr emission from starburst galaxies is due to cosmic rays accelerated by supernovae , the vhegr luminosity above @xmath45 , @xmath97 is then proportional to the star formation rate . for starbursts , this is linearly related to the continuum infrared luminosity @xcite . thus , normalizing by m82 and ngc 253 , we have @xmath231 while this relationship is extremely uncertain ( the normalization varies by a factor of two between m82 and ngc 253 ) , it suggests that ulirgs , or perhaps hyper - luminous infrared galaxies ( hlirgs , @xmath232 ) may be sufficiently bright to contribute to the heating of the igm . while the fluxes from the brightest ulirgs remain much smaller than those associated with the typical tev blazars , there are many more starburst galaxies than agn . at all redshifts ulirgs constitute the high - luminosity tail of the star - forming galaxy luminosity functions @xcite . in the present epoch , the density of ulirgs is roughly @xmath233 , and thus the corresponding vhegr luminosity density of these objects , @xmath234 ( see table 8 of * ? ? ? is negligible in comparison to the tev blazars , roughly @xmath235 . however , due to the steep decline of the luminosity function in the ulirg range , small changes in the location of the break luminosity result in large changes in the comoving luminosity density of ulirgs . as a consequence , the comoving luminosity density of ulirgs grows much faster than that of quasars ( and thus presumably the tev blazars ) nevertheless , even at @xmath27 , roughly the redshift of peak star formation , the comoving number density of ulirgs remains below @xmath236 ( see table 8 of * ? ? ? * ) , corresponding to a comoving vhegr luminosity density of @xmath237 , more than two orders of magnitude smaller than the contemporaneous tev blazar population . for this reason , we neglect the starburst contribution to heating the igm here , though they may represent an secondary source class . stellar - mass objects , such as magnetars and x - ray binaries , generally have difficulty reaching the flux limits required for plasma cooling to dominate the pair beam evolution . even at eddington - limited vhegr luminosities , reaching isotropic - equivalent luminosities of @xmath71 requires beaming factors of roughly @xmath238 , corresponding to jet opening angles of roughly @xmath239 . in practice , the formation of radio jets is believed to be associated with substantially sub - eddington accretion flows in these objects , and thus exacerbating the beaming requirement . more importantly , these sources may suffer compactness problems : the vhegr emission regions in stellar - mass jets must necessarily be very far from the central object to avoid in situ pair - production . finally , were x - ray binaries and magnetars generally strong , persistent vhegr emitters with sufficiently large fluxes , we would expect that many would have been already detected . gamma - ray bursts ( grbs ) are natural candidates due to their large luminosities and strong inferred beaming , and we consider them as a example of the general class of energetic transient sources . unfortunately , little is known about the vhegr emission of grbs . presently there is a single report of a tev signal associated with a grb ( grb 970417a , * ? ? ? * ; * ? ? ? * ) , though due to the large distances at which they can be observed and comparative rarity this is not unexpected . however , such high - energy emission is possible in principle , presumably due to inverse - comptonization of the prompt emission and/or x - ray afterglow ( see section viii of * ? ? ? * and references therein ) . fermi observations of grbs have shown that for many events the band spectrum can be extended to @xmath57 @xcite , though in at least one case a spectral break below @xmath240 has been observed ( grb 090926a , * ? ? ? thus it remains unclear if in practice the high - energy emission is attenuated within the emission region . moreover , since grbs are inherently short - lived events , the luminosity limits described in paper i , which require the vhegr emitting phase to last for a plasma cooling timescale ( roughly @xmath241@xmath242 ) , are not directly applicable . a similar analysis , obtained by limiting the beam growth time to the grb duration , gives vhegr isotropic - equivalent energies of @xmath243 . this is comparable to the total prompt and afterglow emission for only the brightest bursts , comprising roughly 5% of grbs observed by swift @xcite . nevertheless , even assuming that all grbs produce the requisite high energy emission , at an optimistic present local rate of roughly @xmath244 , produces a comoving luminosity density of @xmath245 , roughly three orders of magnitude less than that due to tev blazars at @xmath44 . the previous sections have shown that the vhegr emission from luminous tev blazars heats the igm , quantified the magnitude and stochasticity of this heating locally , and estimated its evolution as a function of redshift . we are now in a position to discuss its impact on the thermal history of the igm in detail . in the following we will show that tev blazar heating can be substantial , dominating late time photoheating , and that its uniform nature naturally imprints its signature onto the temperature - density relation of the igm . the canonical history of the igm is shaped by two important events : h reionization by the first stars at @xmath200@xmath246 and heii reionization by quasars at @xmath247 ( see * ? ? ? * and references therein ) . hydrogen and heii reionization both heated the igm to a @xmath248@xmath249 or higher ( in the case of heii ) . subsequently , the universe cooled via adiabatic expansion , balanced by continuing photoheating due to ionization of recombining h. the entire canonical history of the igm can thus be summarized as a competition between photoheating and adiabatic cooling , punctuated by intervals of sudden photoheating . the addition of tev blazar heating adds an additional extended heating component , and fundamentally alters the canonical picture for the thermal history of the igm . in the following we explore the consequences of blazar heating using the one - zone model originally due to ( * ? ? ? * hereafter hg97 ) ( see also * ? ? ? we begin by introducing this model in detail ( section [ sec : model ] ) , describe thermodynamic consequences of the new heating contribution from blazars and relate this to high-@xmath21 measurements ( section [ sec : igm blazar heating ] ) , and close with a discussion of the implications for the local forest ( section [ sec : lyalocal ] ) . the thermal evolution of a fluid element in the igm is governed by @xmath250 where @xmath251 is the redshift dependent hubble function , @xmath69 is the mass overdensity , @xmath252 is the _ proper _ number fraction of species @xmath253 , relative to the proper number density of baryons , @xmath254 , @xmath255 is the critical mass density of the universe , and @xmath256 is the heating and cooling rate of the gas . the heating and cooling of igm gas is governed by four processes : adiabatic cooling / heating from hubble expansion / gravitational collapse , h / he photoionization heating , h / he recombination cooling , compton cooling , and heating from tev blazars . the evolution of the proper number fraction of the various species is given by @xmath257 where the @xmath258 are the associated atomic rates , not to be confused with the beam instability cooling rates discussed in section [ sec : pair propagation ] . finally , we demand a prescription for the density evolution , i.e. , the evolution of @xmath69 . for this , we follow hg97 and assume the zeldovich approximation : @xmath259 where @xmath260 is the linear growth factor @xcite . the @xmath261 matrix , @xmath262 , is determined by initial conditions . the exact form of this matrix is irrelevant . what is important , however , is the probability distribution of the eigenvalues of this matrix . for a gaussian random field , the solution is known @xcite ; we use the formulation of @xcite . reionization of the universe occurs in two stages : first h is reionized at some large redshift by stars and later he is reionized around a redshift of @xmath263 by quasars ( see , e.g. , * ? ? ? * ) . to model this reionizing history , we adopt a sudden h photoionizing model ( hg97 ) : @xmath264 where @xmath265 is the redshift of h reionization and @xmath266 is the normalization of the reionizing radiation . while , this is not a realistic model of how h reionization occurs , its late time evolution ( especially after @xmath267 ) should be reasonably accurate . this is because the photoionizing background is observed to be roughly constants at these redshifts @xcite and the late time temperature asymptotes to a single value . this `` loss of memory '' of the specific reionization history in the evolution of the igm is typical of reionization calculations @xcite . to model the redshift dependence of h and he photoionization , we use the following spectral model for the radiation : @xmath268 where @xmath34 is the energy of the photon , @xmath269 , @xmath270 , @xmath271 are the threshold energies corresponding to the ionization of hi , hei , and heii , @xmath272 is the redshift of he reionization . the spectral index of @xmath273 is typical of quasars and the spectral model of equation ( [ eq : spectral ] ) is similar to the he reionization model studied by @xcite . our model differs from theirs in that we ignore the density dependent effects of he reionization , i.e. , that dense regions are reionized first . however , since we are interested solely in the magnitude of the effect of blazar heating , adopting this simplified model is justified . the normalization of the photoionizing background in our model is fixed , determined by the h photoionizing rate : @xmath274 where @xmath275 is the photoionizing cross section of hi . we choose the normalization of @xmath276 to be @xmath277 , which is inferred ( with significant uncertainty ) from the mean absorption of the forest ( see for instance * ? ? ? * ; * ? ? ? patch of the igm for the numerical solution ( solid lines ) of equations ( [ eq : t evolution ] ) , ( [ eq : x evolution ] ) , and ( [ eq : delta evolution ] ) using the prescription for the microphysics as specified in the appendix of hg97 . the solid curves are for sudden reionization histories for h and he ( equations ( [ eq : sudden ] ) and ( [ eq : spectral ] ) ) for @xmath278 and @xmath222 and @xmath279 going from left to right . the dashed ( dashed - dotted ) lines show the evolution using the blazar luminosity density , i.e. , using the quasar luminosity density from @xcite to normalize the local heating rate in our standard ( optimistic ) models ( see figure [ f : qlf eos ] and surrounding discussion ) . ] equations ( [ eq : t evolution][eq : spectral ] ) constitute a complete model for the evolution of a fluid element in the igm . we numerically integrate these equations using a prescription for the heating and cooling microphysics specified in the appendix of hg97 . in figure [ f : thermal history ] , we plot the evolution of the temperature for @xmath280 patch with @xmath281 and @xmath222 . the solid lines are the purely photoionized models without the effect of additional heating . for each of these models , we also set the redshift of heii reionization at @xmath282 , which results in a temperature jump at that redshift . we note that at late redshift , the three different ( purely photoionized ) reionization histories asymptote to a single temperature evolution , highlighting the `` loss of memory '' property that is generic to photoionization - dominated models @xcite . realizations at @xmath283 ( black dots ) , @xmath27 ( blue squares ) , @xmath44 ( green triangles ) , and @xmath284 ( red diamonds ) when heating from tev blazars is ignored . the shape of the temperature - density plot can be clearly be fit with a power law ( with a positive index ) and steadily falls from @xmath285 , i.e. , after he reionization . the correspondence between the @xmath286 and @xmath27 points are simply an accident of choosing he ii reionization at @xmath285 and plotting the temperatures and densities at @xmath287 and @xmath220 . namely the universe adiabatically cools , but the injection of heat at @xmath285 resets the temperature and it proceeds cooling from that point onward . ] generally , the thermal history of a given patch depends upon the particulars of the @xmath69-evolution of the patch . in figure [ f : eos ] we show the temperature of @xmath288 realizations of an evolving patch as a scatter plot at a number of redshifts , ranging from @xmath284 to @xmath286 . from this it is clear that the temperature - density relation is well approximated by a power law , consistent with those of hg97 . figure [ f : eos ] represents a typical temperature - density relation , i.e. , @xmath289 relationship , that is typical of most reionization calculations . low density regions in the universe are cooler compared to high - density regions due to decreased recombination ( and hence photoheating ) and a more rapid expansion ( and hence greater adiabatic cooling ) . missing from this simple picture are the effects of shocks and outflows from galaxies , i.e. , feedback ( see @xcite for instance ) . in addition , the temperature of the igm is relatively cool ( @xmath290 k ) for @xmath291 . this is because the temperature is suddenly raised to a few @xmath292 k after h and he reionization , but rapidly cools due to the effects of adiabatic expansion . these generic characteristics are typical of most reionization models and are expected following the arguments of hg97 and @xcite . when the effects of heating due to the vhegr emission from blazars are included , the properties of the igm are substantially altered . the consequence of tev blazar emission is shown for a @xmath293 patch of the igm by the dashed line in figure [ f : thermal history ] . the thermal history begins to deviate significantly from that due to photonionization and adiabatic expansion alone by @xmath294 , becomes dominant near @xmath295 , and peaks at roughly @xmath220@xmath296 at @xmath297 before the rapid decline in @xmath298 combined with adiabatic cooling causes the temperature to fall off ( the range corresponds to the uncertainty in estimating the number of blazars contributing to the heating rate ) . thus , it is clear that heating by blazars is significant , dominates at low redshifts ( following he reionization ) , and potentially dominates the thermal evolution of the igm in low - density regions . the effect of tev blazar heating qualitatively changes the picture of the igm . first , the temperature - density relation is inverted with the low - density regions being the hottest . second , the overall temperature of the igm is significantly hotter . the reasons for both of these are twofold . first , tev blazars are a substantial reservoir of heating , potentially increasing the igm temperature by a few @xmath299 , and dominating the contribution from ionizing photons for @xmath26 . second , the heating rate is nearly independent of density , depending most strongly upon the number density of tev blazars and the number density of uv photons , both of which are nearly uniform ( though see section [ sec : homogeneity ] ) . the effect of a uniform heating rate is that the energy deposited per baryon is substantially larger in more tenuous regions of the universe , with underdense regions experiencing larger temperature increases as a result . ) ) are denoted by gray lines for the standard and optimistic blazar heating models . we use a @xmath280 fluid element in the igm and calculate the amount of heat per baryon ( in units of ev ) per gyr . the sharp jump at @xmath285 is due to heii reionization . aside from this point , it is clear that blazar heating dominates the heating of the igm . at late times @xmath23 , it is larger by over a factor of @xmath246 , though this is substantially greater for underdense regions . ] the dominance of tev blazar heating over photoheating is shown explicitly in figure [ f : rates ] , where the blazar heating rate ( dashed line ) is compared to the photoheating rate ( solid lines ) as a function of redshift for a @xmath293 patch of the igm . at @xmath285 , there is a sudden jump in the photoheating rate due to nearly instantaneous heii reionization . following heii reionization , the blazar heating rate is about an order of magnitude larger than that due to photoheating . specifically , for reasons of simplicity , we have ignored the ionizing photons produced by the quasars . following heii reionization , however , this is no longer an issue . ] . because the photoheating rate is @xmath300 , the dominance of tev blazar heating is even more apparent at lower densities . the contribution of tev blazar heating to the thermodynamics of the igm for the standard model appears to be significant around the period of heii reionization . in our model , this is partially a result of our sudden reionization prescription for heii reionization at @xmath285 . however , we can also show with the following order - of - magnitude calculation that the effect of tev blazar heating must begin to be important around the era of he ii reionization that has been observationally constrained to be around @xmath267 . to begin we first show that he reionization finishes around @xmath301 . the _ comoving _ number density of he is @xmath302 where @xmath303 and @xmath304 are the primordial mass fraction and the atomic number of he , respectively . to estimate the comoving density of heii ionizing photons at @xmath267 , we note that the @xmath305 photon comoving density at @xmath285 from quasars in the @xcite qlf is @xmath306 ( e.g. see figure 9 of @xcite ) . as the spectral index of quasars is @xmath273 , this implies that comoving number density of ionizing heii photons is @xmath307 . thus , the total comoving density of heii ionizing photons produced at @xmath285 is @xmath308 before we compare of equation ( [ eq : hephotons ] ) with ( [ eq : nhe ] ) , we note the it takes roughly 2 - 3 he ionizing photons to completely ionize he and that obscuring material around a quasar will remove half of the ionizing flux @xcite . with these efficiency factors in mind , the comoving density of heii ionizing photons at @xmath267 is just large enough to reionize heii . the amount of excess energy per heii ionization is @xmath309 ev for an ionizing radiation spectral index of -1.6 . given that there are @xmath310 ionizing photons per heii reionization , the excess energy dumped into the igm during heii reionization is @xmath311 where @xmath312 ev is the average excess energy dumped per heii ionization , @xmath303 is the primordial mass fraction of he , and @xmath313 is the atomic number of he . to compare this to the energy dumped from tev blazars , we start with the comoving luminosity density of blazars at @xmath285 is @xmath314 , which we estimate from figure [ f : bld ] . over a hubble time at @xmath285 , the amount of energy dumped into the igm by tev blazars per comoving volume is then @xmath315 which compares favorably with equation ( [ eq : qheii ] ) and shows that as heii reionization is being completed , the effect of blazar heating begins to be pronounced . for convenience , we provide a fitted third - order polynomial formula for our estimated _ proper _ blazar heating rate for @xmath316 : @xmath317 here , @xmath318 for the `` standard '' and `` optimistic '' ( see below ) blazar heating model , respectively . the fits are shown by the solid ( gray ) lines in figure [ f : rates ] . note that we have calculated the heating rate as a heating rate per baryon , i.e. , @xmath319 , so that it is independent of the ionization fraction . + the effect of blazar heating upon the void temperature - density relation is shown in the left panels of figure [ f : qlf eos ] for the `` standard '' ( upper ) and `` optimistic '' ( lower ) heating models . in comparison to the case in which blazar heating is neglected ( fig . [ f : eos ] ) , the voids are hotter at all redshifts . overall , the igm never drops below @xmath320 , with low - density regions remaining substantially hotter than this . by contrast , the temperature of the igm without blazar heating does not _ exceed _ @xmath238 k ( fig . [ f : eos ] ) . in these low - density regions the temperature can exceed @xmath321 with tev blazar heating , almost two orders of magnitude hotter than anticipated by photoheating and adiabatic cooling alone . to show this result more clearly , we show in the right panels of figure [ f : qlf eos ] the ratio between the temperature for a photoheating only model ( @xmath322 ) and for a model including blazar heating ( @xmath323 ) for the `` standard '' ( upper ) and `` optimistic '' ( lower ) heating models . this shows that the temperature in low - density patches in the universe , i.e. , small @xmath324 , increase by a factor of @xmath325 100 and @xmath325 200 for the `` standard '' and `` optimistic '' heating models , respectively . figure [ f : qlf eos ] shows that blazar heating qualitatively changes the igm temperature - density relation . in particular , the igm temperature - density relation is inverted for @xmath326 . this compares favorably with recent evidence for just such an inverted temperature - density relation @xcite . notably , in the lower panel of figure [ f : qlf eos ] we plot as a solid ( gray ) line the temperature - density relation @xmath327 , inferred at @xmath283 empirically by @xcite . here to produce a better match to the empirical measurements of @xcite , we introduce the `` optimistic '' heating model , i.e. , we choose a value of the remaining coefficient of the sky incompleteness correction for tev blazars of @xmath328 , which includes variations of the tev blazar redshift evolution , additional tev sources that may contribute to the plasma instability heating , and potentially spectral variability . as a result , the local heating rate is enhanced over the `` standard '' heating rate by a factor of two or @xmath329 . the match produced between the empirical relation of @xcite and our `` optimistic '' tev blazar model is striking in that it matches the slope in the absence of any tuning as well as the normalization within the uncertainties of our estimated incompleteness correction ( see figure [ f : bld ] ) . this inversion of the temperature - density relation is difficult to reproduce using heii reionization alone @xcite though is a natural consequence of tev blazar emission . energetically , the magnitude of the impact from tev blazars in low - density regions is somewhat surprising . the radiative output from stars and quasars far outstrips that from vhegr sources , yet in practice the heating rate from blazars is much larger . this is because the photoheating rate is ultimately limited by the hii recombination rate . indeed , this is precisely the property invoked to show that the effect of the epochs of h and heii reionization is washed out at low redshift , unless these occurred recently @xcite . by contrast no such limitation exists for tev blazars , paving the way for them to dominate the heating of the low - redshift universe , a point that we have made already in the introduction . instead , the vhegr emission from blazars is deposited in the igm with order unity efficiency . the rate of heating in this case depends linearly on the radiative output of these tev blazars , and is independent of the atomic physics of the igm . in addition these sources have a constant heating rate per unit volume ( whereas photoheating heats per unit mass ) . hence , the effect of blazar heating , which is already pronounced , is amplified relative to photoheating in low - density regions , leading to an inverted temperature - density relation . therefore , not only is the memory of photoheating is erased in low - density regions , but it is overwritten by the record of blazar heating . following this work , we have recently completed a study of the effects of blazar heating in a more detailed hydrodynamic model of structure formation , reported in a follow - up paper @xcite . using the blazar heating prescription given by equation ( [ eq : fit ] ) , we show that the optical depth of the forest is reproduced using a h photoionization rate of @xmath330 or equivalently using the inferred ionizing background from @xcite . we also confirm that the low density igm again possesses an inverted temperature - density relation ( as shown in figure [ f : qlf eos ] ) . in addition , a detailed comparison between the results of our numerical calculations and observations show that a blazar heated universe matches the one- and two- points statistics of the high redshift ly@xmath95 forest , but also matches the line width distribution . this excellent agreement was achieved using the best estimate on the evolution of the photoionizing background ( without tuning ) and is due to the excess energy injection into the low density igm . while tev blazar heating can with significant success reproduce some of the peculiar properties of recent high-@xmath21 observations , the most dramatic departures from the igm thermal history in the absence of blazar heating occur at @xmath331 . this is a result of the relatively recent nature of quasar activity and the cumulative effect of blazar heating . thus we might anticipate dramatic consequences for the local forest , potentially in conflict with studies of nearby clouds ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . however , this does not occur due to the physical nature of the clouds that produce the local forest . the structures responsible for the local forest are almost certainly associated with significant overdensities as suggested by large - scale hydrodynamic computations ( e.g. , * ? ? ? * ) , the same simulations imply that forest contains significant contributions from regions with @xmath332 . ] . these simulated forest calculations provide a means to relate the empirically inferred hi columns to the properties of the dynamical structures responsible for the clouds . for hi columns of @xmath333@xmath334 , the lowest values for which measurements exist , the simulations find that the local forest is produced primarily by the intergalactic filaments , corresponding to @xmath335@xmath336 ( e.g. , see the bottom - left panel of figure 8 and figure 9 of * ? ? ? we emphasize this point with the following order - of - magnitude estimate . an upper limit upon cloud sizes can be obtained using the line widths , @xmath337 , translating into a proper size of @xmath338 @xcite . a lower limit upon the local ionizing background can be inferred from the optical depth of vhegrs , giving @xmath339 @xcite . employing ionization balance to set the hi fraction , for @xmath41 this gives the column densities @xmath340 for @xmath341 . from this it is clear that the nearby clouds must correspond to regions with @xmath342 . at @xmath46 , we would nevertheless expect low - density regions to contribute significantly to the high-@xmath21 forest , as has indeed been found to be the case . ] at these overdensities , the impact of blazar heating on their thermal history is modest at best . by @xmath343 the blazars change the igm temperature by a factor of 2 , raising it to roughly @xmath344 , comparable to the temperatures typical of the local clouds @xcite . at @xmath345 blazar heating is negligible in comparison to photoionization and shock - heating . thus , because low-@xmath21 absorbers are biased toward high - density regions , they serve a poor probes of low - density regions and are unaffected by the effects of tev blazars . the inclusion of blazar heating into these large - scale hydrodynamic calculations is an important next step . however , we do not expect a significant impact on the local forest . a more important impact of tev blazar heating would be on the formation of collapsed structures , which is a topic that we will explore in paper iii . as low - density regions are immune to local probes , we now turn to the question whether there are methods of directly inferring the presence of such high temperatures in these regions . first , we compute the mean comptonization of the cmb due to blazar heating , @xmath346 where @xmath347 is the free electron density . here , @xmath347 is the physical density of free electrons and we integrate along the proper distance of the photons back to recombination , @xmath348 where the minus sign arises due to the choice of coordinates which are centered on the observer . performing the integral with our temperature evolutions at mean density ( i.e. neglecting gravitational heating by formation shocks ) yields mean comptonization values of @xmath349 for our models without blazar heating and those with standard and optimistic blazar heating . to date the best limits on the mean comptonization come from the cosmic background explorer far - infrared absolute spectrophotometer experiment ( cobe firas ) which measure the difference between the cmb and a perfect black - body spectrum @xcite . their upper limit of @xmath350 ( 95% confidence level ) is perfectly consistent with our inferred mean comptonizations . with current technology , it appears to be quite feasible to measure the deviation of the cmb spectrum from a perfect blackbody form with an accuracy and precision of 1 ppm yielding constraints on the cosmic @xmath351-parameter at the level of @xmath352 and provide a spectrum of the anisotropy to 10% @xcite . we note , however , that the mean comptonization is expected to be dominated by gravitational heating which leads to values of @xmath353 inferred from cosmological simulations by @xcite . the mean temperature of these cosmological simulations of @xmath354 suggests that the signal is dominated by collapsed galaxy groups and unvirialized infall regions onto galaxy clusters that constitute the hot component of the warm - hot intergalactic medium . hence , in order to measure the blazar heating signal in the mean comptonization of the cmb , the fluctuating part due to gravitational heating would have to be subtracted first . since galaxy groups dominate the fluctuation power on angular scales of @xmath355 , the measurement of the deviation of the cmb spectrum would have to be performed on these small angular scales which appears to be difficult . second , we estimate the emission of the igm due to free - free emission ( bremsstrahlung ) and synchrotron emission . turning to the question of synchrotron emission , we estimate the synchrotron frequency using a maximal igmf strength of @xmath356 g. at temperatures of @xmath357 k , the electrons in the igm are nonrelativistic . hence the synchrotron frequency is roughly the larmor frequency , @xmath358 . these low frequencies are well below the plasma frequency of the igm , @xmath359 hz and , hence , will be absorbed by the igm . free - free emission from the igm can cause a distortion in the cmb at low frequencies @xcite . as the free - free emissivity scales like @xmath360 for fully ionized primordial gas , it scales weakly with temperature but strongly with clumpiness . as a result , overdense haloes dominate free - free emission , with the contribution from low density igm being 1 - 3 orders of magnitude smaller @xcite . the effect of blazar heating will further suppress this free - free emission ( albeit mildly ) from the igm compared to that of overdense halos . finally , the best constraints on this free - free optical depth of @xmath361 @xcite is still too large compared to optimistic models of free - free emission from overdense haloes by at least an order of magnitude and by @xmath310 orders of magnitude from the smooth igm @xcite . thus far , we have concerned ourselves with the impact of tev blazars on the universe at large . in this section , we invert this argument and discuss how the universe at large can be used as a probe of the physics of vhegr photons and the global properties of tev blazars . namely , the estimated heating rates used in this paper standard and optimistic suffer from a number of uncertainties , including statistical fluctuations in the number of presently observed vhegr sources , the redshift evolution of the tev blazar luminosity density , and the detailed form of the vhegr and ebl spectra . however , due to the high efficiency with which the vhegr luminosity of blazars is converted to heat within cosmic voids , the temperature history of the voids themselves provides a way to determine the cumulative tev blazar luminosity density empirically . this is possible due to a number of fortuitous properties of voids and tev blazar heating . first , the high efficiency with which the vhegr emission of blazars is converted to heat implies that even in voids the universe acts as a calorimeter . second , blazar heating dominates the thermal history of voids for @xmath211 ( see , e.g. , figure [ f : rates ] ) so that this calorimeter is uncontaminated . we caution that heii reionization complicates matters somewhat , though in our model makes a comparatively small correction to the temperature evolution of low - density regions ( see figure [ f : thermal history ] ) . however , this depends upon the particular manner in which heii reionization occurred , and could in principle provide a somewhat larger contribution in inhomogeneous reionization scenarios ( see section 3 of * ? ? ? * for a detailed discussion ) . finally , within voids the adiabatic losses to hubble expansion are very accurately modeled in the linear regime , substantially simplifying the interpretation of void temperature histories . perhaps the greatest uncertainty is the physics of the mechanism which converts pair beams from tev blazars to thermal energy in the igm . in the context of the `` oblique '' instability , what is unclear is the luminosity cutoff of tev blazars probed in this manner , which arises from the competition between inverse - compton and plasma processes . if this cutoff may be conservatively assumed to lie at an isotropic - equivalent @xmath362 , corresponding to a true luminosity 23 orders of magnitude lower due to the presumed jet beaming , then this is dimmer than all but two of the tev blazars known . assuming that the tev blazar luminosity function follows that of quasars , the luminosity density is dominated by sources near the break luminosity , and thus as long as the lower - luminosity cutoff is sufficiently low ( below that of the break luminosity ) it may be neglected . for the tev blazars listed in table [ tab : tevsources ] , this break occurs near @xmath363 , and thus is well above the relevant cutoff . measurements of the thermal history of voids directly corresponds to the tev blazar bolometric luminosity evolution . in this manner , the thermal history of cosmic voids provides an analogous argument to that by @xcite for determining the blazar luminosity density , @xmath298 . such a constraint on the history of @xmath298 can be used to study the history of accretion in the universe and the jet forming efficiency by comparing it to the history quasar or active galaxy luminosity density . while probes of the low - density igm at low redshift are sparse , the situation at @xmath364 is much more hopeful as we have discussed in section [ sec : igm blazar heating ] . thus , precision measurements of the forest at @xmath364 ( e.g. * ? ? ? * ) alongside detailed studies of heii reionization ( * ? ? ? * e.g. ) offer the best constraint both on the fate of vhegr photons and the evolution of tev blazars , i.e. , the blazar luminosity density . in this work , we have explored the effect of tev blazar heating on the thermal history of the igm . we have argued that vhegrs that are sufficiently hard to pair produce off of the ebl , will inevitably dump the majority of their energy into the igm via plasma instabilities . by collating the nearby 28 tev blazars with firm spectral measurements , we have determined the local observed heating rate after correcting for the various selection effects using observations of tev blazars . this local heating rate of @xmath365 , which we call the standard model , can be extended to higher redshift by normalizing it to the @xcite quasar luminosity density . this follows from the important result of paper i , which shows that the local observed blazar luminosity function is well in line with the local quasar luminosity function corrected by a factor of @xmath366 . this tev blazar heating should be relatively homogeneous at all redshifts @xmath163 with greater spatial variations for at higher @xmath21 , approaching order unity at @xmath2 . this redshift dependent blazar heating is substantial and is larger than the photoheating rate by a factor of 15 ( standard ) - 30 ( optimistic ) after he reionization for a @xmath280 patch of the universe . using a simple one - zone model of the igm , we demonstrate that the effect of including tev blazar heating versus not including tev blazar heating leads to qualitative and quantitative changes in the thermal history of the igm . first , the injection of heat into the igm by blazars substantially increases the temperature of the igm . in the case with blazar heating , the temperature of the igm stays above @xmath238 k , with some regions approaching @xmath367 k , whereas without blazar heating , the temperature of the igm tends to stay below @xmath238 k. second , the even volumetric heating rate of blazars impacts the thermal history of low - density regions much more strongly than higher density regions . low - density regions are substantially hotter as a result with temperatures in excess of @xmath368 k. higher density regions on the other hand are not heated as much by blazars . this naturally produces an inverted temperature - density relation which matches the empirical results of @xcite if we increase the amount of blazar heating , i.e. , the optimistic model , to @xmath369 . it also provides an encouraging endorsement of our model as such inverted temperature - density relations are difficult to produce in standard reionization histories . we have demonstrated these salient points more explicitly in a follow up paper @xcite , where we calculate the effect of blazar heating in a hydrodynamic realization of the universe . we show that the comparison between a blazar heated universe and observation of the high redshift ly@xmath95 forest gives excellent quantitative agreement . as our model predicts a substantially hotter low - density igm that standard models predict , we then proceeded to investigate either if this model breaks current constraints on the local temperature of the igm or can be directly measured . unfortunately , the local forest is an ineffective probe of this environment compared to the high-@xmath21 forest as the regions that give rise to the local forest are dense regions that remain relatively unaffected by the effects of blazar heating . other means of directly probing this hot igm via comptonization of the cmb and free - free emission emission are also unlikely . finally , we note that the thermodynamics of the igm can be used as a calorimeter for vhegr emission of blazars in the universe . namely , because the low - density igm is so sensitive to the total amount of energy dumped into it , which is dominated by tev blazars , we argue that the thermal history of the low - density igm can be used to measure the total energy output in vhegrs over cosmic time . in principle , this would allow a determination of the blazar luminosity density as a function of redshift , as well as constrain the history and physics of accretion onto supermassive black holes , i.e. , rates of radiative versus radiative inefficient accretion and jet formation efficiency . however , the contaminating effects of he ii reionization would have to be explored in detail before such physics can be elucidated . we thank tom abel , marco ajello , marcelo alvarez , arif babul , roger blandford , james bolton , mike boylan - kolchin , luigi costamante , andrei gruzinov , peter goldreich , martin haehnelt , andrey kravtsov , ue - li pen , ewald puchwein , volker springel , chris thompson , matteo viel , marc voit , and risa wechsler for useful discussions . we also thank the referee , david weinberg , for a thorough reading of the manuscript and for his constructive comments . we are indebted to peng oh for his encouragement and useful suggestions . we thank steve furlanetto for kindly providing technical expertise . these computations were performed on the sunnyvale cluster at cita . a.e.b . and p.c are supported by cita . gratefully acknowledges the support of the beatrice d. tremaine fellowship . c.p . gratefully acknowledges financial support of the klaus tschira foundation and would furthermore like to thank kitp for their hospitality during the galaxy cluster workshop . this research was supported in part by the national science foundation under grant no . nsf phy05 - 51164 . here we provide a more detailed discussion of how `` likely '' it is for a patch of the universe at a given redshift to experience a large deviation from the average tev blazar heating rate . we will pursue this primarily by attempting to compute the `` number of blazars a patch sees '' as a function of redshift , @xmath370 . this is , however , a poorly defined quantity , the primary difficulty being the determination of which objects to count . in principle , we would like to take a census of those sources `` responsible for the bulk of the heating '' ; in practice this is an ambiguous definition . thus , here we will describe and contrast a number of possible definitions . in the process , we will also elucidate which objects dominate the heating . with the addition of the cyan , short - dash - dotted line , showing the instantaneous number of blazars within a comoving volume of radius @xmath371 , and the green , long - dash - dotted line , showing the number of blazars seen at a particular redshift above a flux limit of @xmath372 , roughly that inferred from the very - high energy gamma - ray observations . the remaining lines are as defined in figure [ f : numblazars ] : the blue , short - dashed line shows the instantaneous number of blazars with intrinsic isotropic - equivalent luminosities above @xmath153 within the @xmath201 surface , the red , dotted line shows the number of blazars above the median luminosity at each redshift within the @xmath201 surface , and the black , solid ( long - dashed ) line shows the number of blazars with individual heating rates that exceed that above which half ( 0.75 times ) of the heating is produced . for reference , our estimate of the sky - completeness corrected number of tev blazars _ observed in the tev _ is shown by the black circle , with the error bars denoting the poisson uncertainty only . note that for all calculations we have assumed @xmath187 and adopted @xmath183 . ] in the interest of completeness we will forgo the assumption that the heating is local , and perform the appropriate cosmological calculations . however , since we ultimately would like to identify @xmath207 as a function of observer redshift , @xmath373 , we require observed - dependent definitions of the standard compliment of cosmological distances . specifically , we use the following generalizations of the standard proper , angular diameter , and luminosity distances , @xmath374 where @xmath375 and @xmath376 correspond to the @xmath205 comoving and proper distances , respectively . note that these reduce to their expected expressions when @xmath214 . while we have already defined the mean free path of high - energy gamma rays in equation ( [ eq : dpp ] ) , we also require a definition of the optical depth that a gamma ray that originates at @xmath21 with _ observed _ energy @xmath34 at @xmath373 accrues during its propagation : @xmath377 } \frac{d d_p}{dz } dz\ , . \label{eq : tau}\ ] ] as with @xmath378 , this differs from the definition of @xmath379 at @xmath214 given in paper i , where there we set @xmath34 to the _ emitted _ energy of the gamma ray . we define the flux , @xmath380 , to be that integrated between a given energy range , @xmath381 to @xmath382 ( usually 100 gev to 10 tev ) , @xmath383 where the observed photon number flux is given by equation ( [ eq : spectra ] ) . for a source at redshift @xmath21 , this corresponds to an absorption corrected flux of @xmath384 and therefore an intrinsic luminosity between energies @xmath385 and @xmath386 of @xmath387 where @xmath388 and @xmath389 are the energies bounding the relevant range at @xmath21 . since we would like to compare luminosities within a band across redshifts ( i.e. , keep @xmath381 and @xmath382 fixed ) , we must correct for the fixed spectral shift induced by the different redshifts . assuming that the spectrum is a power law ( as we shall do in all cases here ) , this implies an additional redshift factor : @xmath390 where @xmath391 is the spectrally average optical depth : @xmath392\ ] ] note that this is closely related to @xmath393 . finally , we will adopt the form of the physical density of tev blazars , @xmath394 , described in the main text assuming a cutoff at @xmath187 , consistent with physical models of high - energy gamma ray blazars ( see , e.g. , * ? ? ? associated with this we have a generalized number density , @xmath395 and luminosity density , @xmath396 defined within a given intrinsic luminosity range . we now present various ways to define the number of high - energy gamma - ray blazars that are relevant for heating . these are compared in figure [ f : nbs ] , which supplements figure [ f : numblazars ] with additional approximations for @xmath207 . our first definition is the simplest one can imagine ; choose an intrinsic luminosity range and use the local density to determine the number within a volume defined by the spectrally - average mean free path . that is , @xmath397 this is an approximation of the number of sources with intrinsic luminosities in the specified range within an approximation of a single optical depth . this is shown for @xmath398 by the cyan dash - dotted line in figure [ f : nbs ] . typically , where the blazar population is rapidly evolving it tends to be a poor estimate of the number of sources , overestimating this number by nearly an order of magnitude at @xmath399 . as long as @xmath224 evolves slowly and @xmath400 it is unnecessary to perform the relevant redshift integral to get the volume element . however , this is not always the case , and thus a more accurate estimate is obtained by integrating the blazar number density within the volume specified by unit optical depth . that is , @xmath401 where @xmath402 is defined implicitly by @xmath403 . for this definition , we may choose @xmath404 in a variety of ways . in figure [ f : nbs ] we show two in particular : that from setting @xmath398 ( blue short - dashed line ) , which may be directly compared with the case shown for @xmath405 , and setting @xmath406 as defined by equation ( [ eq : l0.5 ] ) , i.e. , the luminosity - weighted median luminosity , above which sources produce half of the total local luminosity density ( red dotted line ) . for @xmath407 this approximation for @xmath207 gives similar results to @xmath405 for a fixed @xmath404 . at high redshifts , where @xmath170 is rapidly decreasing , the two diverge substantially . the difference between the @xmath408 and the previous two is more striking , and a consequence for the assumed evolving luminosity distribution of blazars , which is clearly important . while a fixed intrinsic luminosity limit may be useful conceptually , even idealized surveys do not directly probe such a population . instead , most surveys are flux limited , and thus we also define a flux - limited definition of @xmath207 . in this case we set the minimum luminosity via @xmath409 where @xmath410 is a fixed flux limit , from which we obtain @xmath411 \,dz\ , , \label{eq : niii}\end{gathered}\ ] ] this corresponds to the number of objects a flux - limited survey ( in the @xmath45@xmath181 band ) performed by an observer at redshift @xmath373 would detect . as such , it is the most directly comparable to the number of tev blazars that have been observed . this is shown for flux limit comparable to that inferred for the tev sample in table [ tab : tevsources ] , @xmath173 in figure [ f : nbs ] by the green long - dash - dotted line . note that the number of objects found at @xmath205 corresponds nicely to the number observed after correcting for incompleteness of present tev surveys . of course , this is by construction since @xmath170 was obtained from the observed population . nevertheless , it is striking that many more objects would have been observed above this flux limit during earlier epochs , vastly exceeding any of the preceding approximations of @xmath207 . however , it does not necessarily follow that all of these sources will have contributed substantially to the heating rate . the most relevant approximation for @xmath207 , and the one we adopt as our primary definition , is set by the heating rates directly . the general idea is to do what we do naturally at earth : arrange all of the sources visible by an observer at @xmath373 by the local heating rate they induce , from largest to smallest , and count until a fixed fraction of the total heating rate is reached . that is , set @xmath207 to be the minimum number of sources ( on average ) required to produce a given fraction of the total heating rate . given its direct connection with the heating rate , this provides the most natural definition of the number of sources `` responsible for the bulk of the heating . '' to do this , however , we first must explicitly define the heating rate in terms of the appropriate functions . given a fixed spectrum , there is a linear relationship between the local flux and the heating rate , defined by equation ( [ eq : qdot ] ) : @xmath412 } de = \chi(z;z_o ) f(e_m , e_m)\\ & = \chi(z;z_o ) \left(\frac{1+z}{1+z_o}\right)^{2-\alpha } \frac{l}{4\pi d_l^2(z;z_o)}\ , , \end{aligned}\ ] ] where @xmath413 } de \bigg/ \int_{e_m}^{e_m } e^{1-\alpha } de \,,\ ] ] is a function of the shape of the spectrum and the redshifts , similar to @xmath414 . thus , a given heating rate defines an intrinsic luminosity limit for a source at a given redshift : @xmath415 with this , we may compute the heating rate as a function of @xmath416 , i.e. , the heating rate associated with sources that produce a local heating larger than some limit : @xmath417 \,dz\ , . \end{aligned } \label{eq : qdiv}\ ] ] the heating from _ all _ gamma - ray blazars is obtained simply by setting @xmath418 . on the other hand , we may set @xmath416 implicitly via @xmath419 where @xmath420 ranges from 0 to 1 , yielding a heating rate limit at each redshift that we shall call @xmath421 . from this , we may then obtain a number of contributing blazars : @xmath422 \,dz\ , . \label{eq : niv}\end{gathered}\ ] ] this is shown for @xmath423 and @xmath424 in figure [ f : nbs ] by the black solid and long - dashed lines , respectively . while both are similar to the other measures of @xmath207 below @xmath197 , above this redshift they fall much more rapidly . this is due to the shift of @xmath224 towards higher luminosities , and thus the luminosity density ( and therefore heating rate ) becomes dominated by fewer , more luminous sources . nevertheless , @xmath207 is a rapidly increasing function of @xmath420 , as evidence by the fact that @xmath425 increases by approximately an order of magnitude when @xmath420 increases from @xmath216 to @xmath426 . [ [ location - and - properties - of - the - sources - responsible - for - the - bulk - of - the - heating ] ] location and properties of the sources responsible for the bulk of the heating ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ armed with a definition of @xmath207 , we can now address which class of high - energy gamma - ray blazars dominates the heating rate . that is , we can assess whether the local heating is dominated by close , intrinsically dim objects or by distant , intrinsically luminous sources . this is done simply by inspecting the integrands in equations ( [ eq : qdiv ] ) and ( [ eq : niv ] ) . however , to interpret these , we will first build some intuition based upon an extremely simplified model , for which an analytical result is trivially obtained . ) of the sources that contribute to the total number ( top ) and heating rate ( middle ) , and the limiting and effective luminosity ( bottom ) for a flux - limited sample of sources described by a fixed , cutoff power - law luminosity function in a static euclidean universe . different colors correspond to different luminosity function power - laws ( @xmath427 , @xmath428 , @xmath429 , @xmath430 , @xmath431 , and @xmath432 shown in violet , blue , cyan , green , orange , and red ) , with the associated luminosity - weighted luminosity function shown explicitly in the inset . for reference , the radius at which @xmath433 is shown by the vertical dotted line . in comparison to the approximate @xmath404 we used ( black solid ) , we show the luminosity limit when absorption is included by the black dashed line , and the @xmath434 defined in equation ( [ eq : leffe ] ) are shown by the dot - dash lines . this may be compared directly with figure [ f : nbds ] . ] for a flux - limited sample of sources described by a fixed , cutoff power - law luminosity function in a static euclidean universe . different colors correspond to different luminosity function power - laws , with the associated luminosity - weighted luminosity function shown explicitly in the inset . this may be compared directly with figure [ f : dqdl ] . ] we begin with a toy problem in which we consider a fixed power - law luminosity function in a static , euclidean universe . specifically , we choose , @xmath435 where @xmath436 is the heaviside function , @xmath437 is the overall normalization of the luminosity function , @xmath438 is a maximum luminosity ( approximating a break ) , and @xmath439 is an arbitrary constant . we will also assume that @xmath47 is independent of energy , i.e. , there is some characteristic value for which @xmath440 and we may bring it out of the energy integral in the definition of @xmath208 . as a consequence , a fixed limit in @xmath416 corresponds directly to a fixed flux limit , @xmath410 . in a euclidean universe the number of objects takes the particularly simple form , @xmath441 the flux limit , @xmath410 , gives a luminosity limit of @xmath442 , where we have ignored the optical depth . since we will be most concerned with how peaked the various integrands are at nearby distances , this is not a significant oversight ( including it would serve to make them only more so ) . the definition of @xmath404 implies a maximum distance as well , with @xmath443 , and therefore @xmath444 . thus , we have @xmath445\\ & = \frac{4\pi d_*^3 \tilde{\phi } _ * } { \ln10 } \int_0 ^ 1 dx \,\xi^{-1 } x^2 \left ( 1 - x^{2\xi } \right ) e^{-x / x_{\rm pp}}\ , . \end{aligned}\ ] ] from this we trivially obtain @xmath446 providing some notion of the location of the most numerous sources . typical values of @xmath439 range from @xmath447 to @xmath432 , in practice , and @xmath448 is shown in the top panel of figure [ f : euds ] for a variety of choices of @xmath439 within this range . the integral for @xmath449 is similarly simple , @xmath450 and thus , @xmath451 giving an idea of the location of the sources responsible for the bulk of the heating . for a variety of @xmath439 , @xmath452 is shown in the middle panel of figure [ f : euds ] . generally , we find that for all but the largest @xmath439 , @xmath453 and @xmath452 are peaked at small distances . in particular , both are typically dominated by @xmath454 . the bottom panel of figure [ f : euds ] shows the flux - limited @xmath404 with ( dashed ) and without ( solid ) absorption included . including absorption suppress the contributions at large @xmath455 , forcing @xmath453 and @xmath452 to be even more strongly peaked at small distances . since @xmath404 is a strong function of @xmath456 , contributions from different distances have different luminosity distributions . it is possible to roughly characterize this by defining a typical luminosity , @xmath434 , associated with contributions at a given @xmath456 : @xmath457 with @xmath404 , this is shown in the bottom panel of figure [ f : euds ] . generally @xmath434 is larger than @xmath404 , and for small @xmath458 substantially so . thus , even for nearly flat @xmath459 ( @xmath460 ) the objects that contribute to the heating are _ not _ dominated by the numerous , intrinsically dim objects with luminosities @xmath461 . alternatively , we may perform the integral over @xmath456 first , in which the flux limit implies a maximum distance to which a given object can be seen , @xmath462 , providing some insight into luminosity of the sources responsible for the heating . doing so yields @xmath463 this is shown in figure [ f : dqedl ] for a variety of @xmath439 . in all cases the luminosity at the break in @xmath459 , i.e. , @xmath438 , contributes most significantly to the heating rate . however , the relative importance of lower - luminosity objects does depend upon the luminosity function ; flat luminosity functions ( i.e. , @xmath427 ) have many low - luminosity sources , and thus induce more broad @xmath464 . nevertheless , it appears that the heating rate in our simple toy model is generally dominated by nearby objects near the peak in the luminosity function . -limited sample of objects . different colors correspond to different @xmath373 , indicated by the left - most value of @xmath21 for which each intersects the horizontal axis ( @xmath373 ranging from 0 to 4 ) , and are consistent with those used in figure [ f : dqdl ] . for reference , the redshifts at which @xmath465 and @xmath186 are shown by the vertical and horizontal dotted lines , respectively . in comparison to @xmath404 , we also show @xmath434 ( as defined by equation ( [ eq : leffb ] ) ) by the dash - dot lines . finally , the upper - right panel shows @xmath466 for each of the redshifts for which distributions are shown in the other panels , with corresponding colors ( note that this shows the same relative dynamic range as that in the inset of figures [ f : euds ] and may thus be directly compared ) . ] we now return to the physically relevant case : heating due to tev blazars with an evolving luminosity function in an evolving universe . here we specify the distributions of the sources responsible for producing a fraction @xmath420 of the total heating rate . in this case , we may immediately read off @xmath467 and @xmath468 , the analogs of @xmath469 and @xmath470 , from equations ( [ eq : niv ] ) and ( [ eq : qdiv ] ) , respectively , yielding , @xmath471\\ \begin{aligned } \frac{d\dot{q}}{dz}(z;z_o ) & = 4\pi d_a^2(z;z_o ) \frac{dd_p}{dz } \left(\frac{1+z}{1+z_o}\right)^{2-\alpha}\\ & ~~~\,\quad\times\frac{\chi(z;z_o)}{4\pi d_l^2(z;z_o ) } \tilde{\lambda}_b\left[z;l_m(z;z_o,\dot{q}_{\mathcal{q}}),l_m\right]\ , . \end{aligned } \end{gathered}\ ] ] these are shown , normalized by their integrated values , for @xmath373 ranging from @xmath432@xmath220 in figure [ f : nbds ] for @xmath423 . in addition we show an analogously defined characteristic luminosity , @xmath472 } { \tilde{\phi}_b\left[z;l_m(z;z_o,\dot{q}_{\mathcal{q}}),l_m\right ] } \ , . \end{aligned } \label{eq : leffb}\ ] ] the generic features of our static toy model are also apparent here . at all observer redshifts the heating rate is dominated by the nearest sources . at low @xmath373 , where @xmath473 is nearly flat , the number of objects is also heavily weighted towards nearby objects , well within the redshift at which @xmath201 . however , @xmath467 and @xmath468 evolve with observer redshift due to both , the intrinsic evolution of @xmath170 and the background universe . as a consequence , by @xmath474 the peak of @xmath467 has moved to the @xmath201 redshift , implying that @xmath402 ( where @xmath475 ) is not a particularly accurate estimate of the redshifts that contribute significantly to the heating . this is further supported by the high-@xmath373 behavior of @xmath467 , which once again is heavily weighted at redshifts inside of @xmath402 due to onset of the decline in the blazar population . the typical luminosities of objects responsible for the heating also evolve . at @xmath476 these are roughly @xmath154 , rising to @xmath477 by @xmath478 . we also compute the heating rate per logarithmic decade in luminosity : @xmath479 ( where @xmath480 is determined implicitly by @xmath481 ) shown in figure [ f : dqdl ] for a number of @xmath373 . at @xmath476 the distribution of heating rates is a relatively broad function of @xmath97 centered near @xmath482 . until @xmath478 , as @xmath373 grows @xmath483 becomes increasingly peaked and centered upon increasingly larger luminosities . above @xmath478 this trend reverses , though the distribution of luminosities that contribute appreciably to the heating rate never becomes comparable to that in the present epoch . thus , generally it appears that the heating is due predominantly to nearby objects with luminosities comparable to @xmath153 . , m. , et al . 2010 , in astronomical society of the pacific conference series , vol . 427 , astronomical society of the pacific conference series , ed . l. maraschi , g. ghisellini , r. della ceca , & f. tavecchio , 302
the universe is opaque to extragalactic very high - energy gamma rays ( vhegrs , @xmath0 ) because they annihilate and pair produce on the extragalactic background light . the resulting ultra - relativistic pairs are commonly assumed to lose energy primarily through inverse compton scattering of cosmic microwave background photons , reprocessing the original emission from tev to gev energies . in ( * ? ? ? * paper i of this three paper series ) , we argued that this is not the case ; powerful plasma instabilities driven by the highly anisotropic nature of the ultra - relativistic pair distribution provide a plausible way to dissipate the kinetic energy of the tev - generated pairs locally , heating the intergalactic medium ( igm ) . here , we explore the effect of this heating upon the thermal history of the igm . we collate the observed extragalactic vhegr sources to determine a local vhegr heating rate . given the pointed nature of vhegr observations , we estimate the correction for the various selection effects using observations of high and intermediate peaked bl lacs . as the extragalactic component of the local vhegr flux is dominated by tev blazars , we then estimate the evolution of the tev blazar luminosity density by tying it to the well - observed quasar luminosity density , and producing a vhegr heating rate as a function of redshift . this heating is relatively homogeneous for @xmath1 , but there is greater spatial variation at higher redshift ( order unity at @xmath2 ) because of the reduced number of blazars that contribute to local heating . we show that this new heating process dominates photoheating in the low - redshift evolution of the igm and calculate the effect of this heating in a one - zone model . as a consequence , the inclusion of tev blazar heating qualitatively and quantitatively changes the structure and history of the igm . due to the homogeneous nature of the extragalactic background light , tev blazars produce a uniform _ volumetric _ heating rate . this heating is sufficient to increase the temperature of the mean density igm by nearly an order of magnitude , and at low densities by substantially more . it also naturally produces the inverted temperature - density relation inferred by recent observations of the high - redshift forest , a feature that is difficult to reconcile with standard reionization models . finally , we close with a discussion on the possibility of detecting this hot low - density igm suggested by our model either directly or indirectly via the local forest , the comptonized cosmic microwave background , or free - free emission , but find that such measurements are currently not feasible .
effective field theories is a subject of great interest in theoretical physics not only due to their potential applications but also because they provide new insights into the way we look at field theories @xcite . from this perspective nonrenormalizable models have acquired a new status as they may become physically relevants at low energies @xcite . the point is that , if the scale of energy one is interested is low enough , the ambiguities due to the virtual states of high energy do not show up or , equivalently , are not meaningful . on the energy interval where this happens the theory proceeds as an usual renormalizable one . nonetheless , as observed in @xcite , the use of a mass independent regularization is almost mandatory to guarantee that high order counterterms can be effectively neglected . the ultraviolet behavior of the green functions may be changed by a rearrangement of the perturbative series . in fact , the incorporation of vacuum polarization effects in general improves the convergence properties of the resumed series ; this mechanism is well known to be operative in the context of the @xmath0 expansion . in particular , gross neveu @xcite or thirring @xcite like four fermion interactions which in ( 2 + 1 ) dimensions are perturbatively nonrenormalizable become renormalizable within the framework of the @xmath0 expansion @xcite . this result has motivated a series of investigations on the properties of these theories @xcite . in particular , using renormalization group ( rg ) methods , it has been proved that the @xmath1 component gross neveu model in 2 + 1 dimensions is infrared stable at low energies but has also a nontrivial ultraviolet stable fixed point . these facts indicate that the theory could be perturbatively investigated if the momentum is low enough . this actually would be the only remaining possibility for small @xmath1 . it has recently been conjectured that in 2 + 1 dimensions , besides the @xmath0 expansion , there is another way to improve the ultraviolet behavior of feynman amplitudes . by coupling fermion fields to a chern - simons field the scale dimension of field operators could be lowered possibly turning non renormalizable interactions into renormalizable or , better , super - renormalizable ones . using a sharp cutoff to regulate divergences , this idea was tested in @xcite where the effect of the cs field over massless self - coupled fermions with a quartic , gross neveu like , interaction was studied . in this communication we will pursue this study further by considering massive fermions and adopting dimensional renormalization @xcite as a tool to render finite the feynman amplitudes . in this way we evade the ambiguity problem associated with the routing of the momentum flowing through the associated feynman graphs @xcite . nevertheless , it should be stressed that our calculations are valid insofar , as said above , the effect of the higher order counterterms can be neglected . otherwise , new couplings should be introduced . our investigations , restricted to the case of fermions of just one flavor , i. e. @xmath2 , show that , differently to what happens for large @xmath1,the renormalization group beta function has only a trivial infrared stable fixed point . moreover , the operator dimensions of the basic field and composites of canonical dimension lower than three are monotonic decreasing functions of the cs parameter . this indicates that the feynman amplitudes have a better ultraviolet behavior if the underlying theory is renormalizable . however , no improvement in the ultraviolet behavior seems to occur if the composite operators have canonical dimension bigger than three . our work is organized as follows . in section ii some basic properties of the model as feynman rules , ultraviolet behavior of feynman diagrams and comments on the regularization procedure are presented . the derivation and calculation of the renormalization group parameters are indicated in section iii . section iv contains a discussion of our results as well as our conclusions . details of the calculations of the pole part of the relevant amplitudes are described in the appendices [ appendixa ] and [ appendixb ] . we consider a self - interacting two component spinor field minimally coupled to a cs field . the lagrangian density is given by @xmath3 the dirac field @xmath4 represents particles and anti - particles of spin up and the same mass @xmath5 ( the parameter m is to be taken positive ) @xcite . the gross - neveu term in ( [ 1 ] ) is the most general lorentz covariant quartic self - interaction , for the thirring - like vector interaction is not independent but satisfies @xmath6 @xmath7 is a gauge fixing parameter but for simplicity we will always work in the landau gauge , formally obtained by letting @xmath8 . in this gauge , the green functions may be computed using the feynman rules depicted in fig . [ fig1 ] . for convenience , we have introduced auxiliary dotted lines , hereafter called auxiliary gn lines , to clarify the structure of the four fermion vertex . divergences show up , the degree of superficial divergence of a generic graph @xmath9 being @xmath10 where @xmath11 and @xmath12 are the number of external lines associated with the propagators for the the chern simons and the fermion fields , respectively ; @xmath13 denotes the number of quartic vertices in @xmath9 . the model is of course nonrenormalizable and the number of counterterms necessary to render the amplitudes finite increases with the order of perturbation but , to a given order , the number of counterterms is finite . to do calculations we will employ dimensional renormalization starting at the space - time dimensio @xmath14 . it is therefore convenient to introduce a dimensionless coupling @xmath15 and a renormalization parameter @xmath16 through @xmath17 and @xmath18 , where @xmath19 must be set zero at the end . the massive parameter @xmath20 must be considered much bigger than any typical momenta and than the fermion mass @xmath5 ; it sets the scale which limits the region where our results are valid . divergences will appear as poles in @xmath21 and a renormalized amplitude is given by the @xmath21 independent term in the laurent expansion of the corresponding regularized integral . however , at one loop level no infinities will remain after the remotion of the regulator . this is so because the poles for a graph @xmath9 may occur only at even values of the degree of superficial divergence of @xmath9@xcite . moreover , it is easy to check that asymptotically , i. e. , for zero external momenta , one loop graphs with even degree of divergence are odd functions of the loop momentum and therefore vanishes , after symmetric integration . as known , green functions of renormalizable models , which have been made finite by the subtraction of pole terms in the dimensionally regularized amplitudes , satisfy a t hooft - weinberg type renormalization group equation @xcite . nonrenormalizable models require special consideration since the form of the effective lagrangian changes with the order of perturbation . however , at sufficient small momenta , such that the effect of the new counterterms may be neglected , the green functions will still approximately satisfy the rg equation . thus , although being nonrenormalizable , for small enough momenta , the green function of the theory ( [ 1 ] ) satisfy the following renormalization group equation @xmath22\gamma^{(n)}(p_1 , \ldots p_n)\approx 0 , \ ] ] where @xmath23 denotes the vertex function of @xmath1 fermion fields ( since @xmath24 is not a dynamical field we shall not consider vertex functions having external vector fields ) . the symbol @xmath25 means equality in the region where all counterterms different from those terms already present in ( [ 1 ] ) can be neglected . as a consequence of the coleman hill theorem , which states that all radiatives corrections to the cs term are finite , @xcite , the @xmath26 function for the cs coupling @xmath27 vanishes identically ; that explains why the term with a derivative with respect to @xmath27 is absent from ( [ 2 ] ) . the coefficients @xmath28 , @xmath29 and @xmath9 in equation ( [ 2 ] ) may be obtained by formally computing the action of the differential operator over the two point and four point green functions . for the two point function , up to second order in the coupling constants , we have @xmath30 where the limit @xmath31 must be understood . in the above expression @xmath32 , @xmath33 , denote the regularized feynman amplitudes . in particular , the graphs ascribed to @xmath34 , @xmath35 and @xmath36 have been depicted in the figures [ fig2 ] , [ fig3 ] and [ fig4 ] , respectively ; @xmath37 is an operator to remove the pole term in the amplitudes to which it is applied . as mentioned earlier , the amplitudes @xmath38 and @xmath39 which are associated with one loop diagrams are finite . inserting ( [ 3 ] ) into ( [ 2 ] ) allows us to determine the coefficients @xmath40 and @xmath9 as follows . initially notice that as @xmath20 enters into the perturbative expansion only in the combination @xmath41 , fixing @xmath29 in lowest order as being equal to @xmath15 eliminates all contributions of the term with the derivative with respect to @xmath20 in ( [ 2 ] ) . after that , up to the order we will study , in ( [ 2 ] ) there will be no mixing of higher order contribution to @xmath29 with those to @xmath9 and @xmath28 . using the expansions , @xmath42 where the sum is restricted to @xmath43 , we get @xmath44 where @xmath45 and @xmath46 , for @xmath47 are defined by writing the pole term for the amplitude @xmath48 as @xmath49 the appendix [ appendixa ] presents a detailed analysis of the various contributions to these parameters . from ( [ i15 ] ) , ( [ i18 ] ) and ( [ i20 ] ) , the final result is @xmath50 letting @xmath51 , we note that our determination of @xmath9 agrees with @xcite . in the region @xmath52 , however , our result presents corrections for nonvanishing fermion mass . to fix @xmath29 we look now at the four point green function , which up to third order is ( we omit contributions which by power counting are finite ) @xmath53 here , again , the one loop amplitudes @xmath54 and @xmath55 are finite because of the use of the dimensional regularization . inserting ( [ 7 ] ) , the expansion @xmath56 , and @xmath28 and @xmath9 given in ( [ 4a ] ) and ( [ 4b ] ) into ( [ 2 ] ) , we obtain @xmath57 , @xmath58 and @xmath59 in the above expressions , @xmath60 for @xmath47 are related to the pole part of the amplitudes @xmath61 through , @xmath62 in the appendix [ appendixb ] we have collected the results of the calculations of the pole part of the relevant graphs . using [ b4 ] we obtain , finally , @xmath63 an inspection of eq . ( [ 8a ] ) shows that the renormalization group _ beta _ function has @xmath64 as a fixed point . as , for @xmath65 , @xmath66 the origin is an infrared stable fixed point . actually , this is the only existing fixed point . here we are in disagreement with ref . @xcite where a line of fixed points was found . the diverse conclusions are perhaps due to the use of different regularizations but a more direct comparison of the methods seems unfeasible as the calculations in @xcite were not spelled out . we will examine now the dimensions of some operators . as seen before the basic field @xmath67 has operator dimension @xmath68 , and so at @xmath64 the green functions of the fermion field have an improved ultraviolet behavior as @xmath27 increases . similar results are obtained if one considers composite operators of canonical dimension less than three . the simplest of them , the mass operator @xmath69 ha an anomalous dimension given by @xmath70 where @xmath71 is the residue coming from graphs contributing to the vertex function with the insertion of the mass operator , @xmath72 and having two external fermionic lines . for practical purpose this residue may be computed by taking the mass derivative of the contributions calculated in the item 1 of appendix [ appendixa ] . the result is @xmath73 . thus the dimension of @xmath74 turns out to be equal to @xmath75 from the computation of the anomalous dimension of the operator @xmath76 one could easily obtain the dimension of @xmath77 . indeed , the dimension of @xmath77 is given by ( [ 8b ] ) but with the replacement of @xmath71 by @xmath78 . thus @xmath79 the determination of the anomalous dimension of the operator @xmath80 at @xmath64 is more complicated due to the fact that renormalization in general produces a mixing with other operators of dimension lower or equal to four . however if we restrict the calculation to the @xmath81 case , as we will do , only operators of dimension four need to be considered . a further simplification is obtained by considering only ( formally ) integrated operators . we have , @xmath82 & = & a_1\int d^3 x ( \bar\psi \psi)^2 + a_2 \int d^3 x \bar \psi \partial^2 \psi\\ \int d^3x n[\bar \psi \partial^2 \psi ] & = & b_1\int d^3 x ( \bar\psi \psi)^2 + b_2 \int d^3 x \bar \psi \partial^2 \psi\end{aligned}\ ] ] where the symbol @xmath1 indicates a normal product prescription corresponding to the subtraction of the pole terms . a direct calculation gives that @xmath83 , @xmath84 and @xmath85 . a straightforward analysis shows now that the dimensions of @xmath86 $ ] and @xmath87 $ ] are given by @xmath88 respectively . one sees that , at least for the operators that we explicitly considered , the operator dimension decreases with @xmath27 accordingly the canonical dimension is lower or equal to three , in accord with @xcite . however , if the canonical dimension is bigger than three , the operator dimension increases with @xmath27 so that no improvement for nonrenormalizable interactions results . in this appendix we shall present a detailed analysis of the contributions to the pole part of the two point vertex function . due to the use of the dimensional regularization the poles only appear at the two loop level , beginning at second order in the coupling constants . we will examine separately each order of perturbation , i. e. , the orders @xmath89 , @xmath90 and @xmath91 , we have , figure [ fig2]@xmath92 @xmath93 \nonumber \\ & \times&\frac{\left[\gamma^{\rho}(\not \ ! p-\not \ ! k_{1}+m)\gamma^{\theta}\right]}{(k^{2}_{2}-m^2)\;[(k_2-k_1)^2-m^2]\;[(p - k_1)^2-m^2]\;(k^{2}_{1})^2 } \nonumber \\ & = & \left(m\;a_3(a)+\not \ ! p\;b_3(a)\right)\frac{1}{\epsilon},\end{aligned}\ ] ] where , on the right hand side of the first equality , we have introduced the operator @xmath37 to extract the pole part of the expression to which it is applied . from ( [ i1 ] ) we obtain for simplicity , the trace was taken in the integrand of ( [ i1 ] ) and calculated directly at @xmath95 . this , of course , does not affect the result for the pole part of the integrals . analogously , @xmath96\;[(p - k^{2}_1)^2-m^2]\;(k^{2}_{1})^2},\end{aligned}\ ] ] where @xmath97 here and in what follows we shall adopt the following procedure for performing the integrals . we first consider the @xmath98 integral and use feynman s trick , @xmath99 } { \gamma[\alpha_1]\gamma[\alpha_2 ] } \int_{0}^{1 } d x \,\frac { x^{\alpha_1 - 1 } ( 1-x)^{\alpha_2 - 1}}{[x a_1 + ( 1-x ) a_2]^ { \alpha_1+\alpha_2}}\ ] ] to reduce the denominators containing the variable of integration to only one denominator . after integrating in @xmath98 we use again feynman s formula ( [ i5 ] ) to combine the denominators that depend on @xmath100 . we then integrate over @xmath100 and , finally , perform the parametric integrations . in the present case , after integrating on @xmath98 , we get @xmath101\end{aligned}\ ] ] @xmath102\;(k^{2}_{1})^{2 } } \nonumber \\ & \times & \left[4\gamma(1-\frac{d}{2})\;\left((d-2)(k_1\cdot p)-k^{2}_1\right)\;\delta^{\frac{d}{2}-1}+8x(1-x ) \right . \nonumber \\ & \times & \left . \gamma(2-\frac{d}{2})\;\left((k_1\cdot p)k^{2}_1-(k^{2}_1)^2\right)\;\delta^{\frac{d}{2}-2}\right],\end{aligned}\ ] ] where @xmath103 . continuing our calculation , we would introduce two new parametric integrations as there are now three different denominators ( we take @xmath104 as a new denominator ) depending on @xmath100 in each of the terms of the above expressions . however , as the result does not depend on @xmath5 and we are looking only for the pole part of the amplitudes , we can speed up the calculation by modifying the dependence on @xmath5 of some denominators . for example , without changing the final result , we can replace the first term on the right hand side of ( [ i6 ] ) by @xmath105 similarly , in the computation of @xmath106 one can set @xmath81 in the expression for @xmath107 so that one has to use only one parametric integral . following this recipe , after integrating in @xmath100 we obtain ( @xmath108 ) , @xmath109 \int_{0}^{1}dx \int_{0}^{1}dy \ , ( 1-y ) \nonumber \\ & \times & [ \frac{(1 - 2 y ) m^2}{a}]^{d-3 } \left [ \frac{y^{-d/2}}{a^{3-d/2 } } + \frac{d}{2 } ( 1 + x - x^2 ) \frac{y^{1-d/2}}{a^{4-d/2}}\right ] \nonumber\\ & = & -\frac{5i}{8 \epsilon}\end{aligned}\ ] ] and @xmath110 \int_{0}^{1}dx \int_{0}^{1}dy \ , \frac{(1-y)^{2-d/2}(y m^2-y(1-y)p^2)^{d-3 } } { [ x(x-1)]^{1-d/2 } } \nonumber \\ & \times & \left [ \frac{(1 - 5 y)\gamma[1-d/2]}{\gamma[3-d/2]}- 2 \frac{\gamma[2-d/2]\gamma[5-d/2]}{\gamma[4-d/2]^2 } ( 1-y ) ( 1- 7 y)\right ] \nonumber \\ & = & - \frac{i}{24 \epsilon}.\end{aligned}\ ] ] figure [ fig2]@xmath111 @xmath112}{[(p - k_{1})^2-m^2]^2\;[(p - k_1-k_2)^2-m^2]\;k^{2}_{1}\;k^{2}_2 } \nonumber \\ & = & \left(m\;a_3(b)+\not \ ! p\;b_3(b)\right)\frac{1}{\epsilon}\end{aligned}\ ] ] following the same steps as in previous case , we obtain @xmath113 figure [ fig2]@xmath114 @xmath115 } { [ ( p - k_{1})^2-m^2]\;[(p - k_1-k_2)^2-m^2]\;[(p - k_{2})^2-m^2]\;k^{2}_{1}\;k^{2}_2 } \nonumber \\ & = & \left(m\;a_3(c)+\not \ ! p\;b_3(c)\right)\frac{1}{\epsilon}.\end{aligned}\ ] ] the computation of this expression is a bit more complicated because one has to introduce three feynman parametric integrals . the final result is , nevertheless , simple , @xmath116 figures [ fig3]@xmath119 and [ fig3]@xmath120 . both diagrams have the structure of a product of two one loop graphs . the corresponding dimensionally regularized amplitudes does not have a pole at @xmath95 . figure [ fig3]@xmath121 . actually , this diagram does not contribute because the corresponding analytic expression is finite . indeed , we have @xmath122 \epsilon_{\mu\nu\rho } } { ( k_{2}^{2}-m^2 ) ( ( k_1+k_2)^2-m^2 ) } \frac { k_{1}^{\rho}}{k_{1}^{2 } } \nonumber\\ & = & - 4 i{\cal t}\frac{\partial \phantom { a}}{\partial m}\int\frac{d^dk_1}{(2\pi)^d}\frac{d^dk_2}{(2\pi)^d}\;\frac{m}{(k_{2}^{2}-m^2 ) ( ( k_1+k_2)^2-m^2)}=0\end{aligned}\ ] ] since the integral in the second equality has the structure of a product of two one loop integrals . figure [ fig3]@xmath125 . we have the following contribution @xmath126\;(k^{2}_2-m^2)\;((k_2-k_1)^2-m^2)\;k^{2}_1}\nonumber\\ & = & -i\frac{(m^2 a_4(f)+ m\not \ ! p b_4(f))}{\epsilon}\end{aligned}\ ] ] from which one obtains @xmath127 \3 . order @xmath91 graphs . there are only the two diagrams shown in fig @xmath128 . we get , @xmath129(k_{2}^{2}-m^{2})[(k_2-k_1)^{2}-m^{2}]}\nonumber\\ & \times & \{(\not \ ! k_2 + m)(\not \!k_2-\not\ ! k_1 + m)- { \rm tr}[(\not \ ! k_2 + m)(\not \!k_2-\not\ ! k_1 + m)]\}\nonumber \\ & = & -\frac{(m^3 a_5 + m^2 \not \ ! p b_5)}{\epsilon},\end{aligned}\ ] ] in this appendix we will discuss the calculation of the pole part of the four point vertex function which is needed for fixing the renormalization group beta function . actually , since all we need is the constant part of the residue , the calculation of the relevant graphs will be done at zero external momenta . the first observation is that , up to the order we are interested , i. e. , third order , there are too many graphs . to be systematic , we will separate them accordingly they have or have not closed fermionic loops . moreover , if they do not posses fermionic loops we group them accordingly the number of cs or auxiliary gn lines linking the two fermion lines crossing the diagram . many diagrams cancel because of furry s theorem ; this is the case if there is a fermionic loop with an odd number of attached cs lines . other diagrams have the structure of a product of two one loop graphs and therefore are finite . we shall not consider these two types of graphs any longer . the anti - symmetrized amplitude for a graph @xmath9 has the generic structure of a product , @xmath131 , where a and b refers to the propagators and vertices associated to the two fermion lines and c to the others factors ( @xmath132 indicates the anti - symmetrized direct product ) . using this notation , one can verify that @xmath133\,\,{\rm tr}[b]-{\rm tr}[ab])c,\ ] ] for example , from the analytic expression for the graph shown in fig . [ fig5 ] , @xmath134\otimes [ \gamma^\nu s_f(-k_2 ) ] { \rm tr}[s_f(k_1 ) s_f(k_1-k_2 ) ] \epsilon_{\mu\nu\lambda } \frac{k_{2}^{\lambda}}{k^{2}_{2 } } , \end{aligned}\ ] ] [ b2 ] where @xmath7 is a combinatorial factor , we determine @xmath135 , @xmath136 and @xmath137 as @xmath138\\ b&= & [ \gamma^\nu s_f(-k_2)]\\ c&= & \lambda\epsilon_{\mu\nu\lambda}\frac{k_{2}^{\lambda}}{k^{2}_{2}}\end{aligned}\ ] ] the results for the pole parts are summarized in the tables @xmath135 and @xmath136 , which corresponds to the two cases mentioned above . in the table a are listed the results from graphs with one closed fermionic loop ; these have been arranged accordingly the number of cs vertices in the loop . table [ appendixb ] exhibits the pole part of graphs without fermionic loop . they have been collected into types ( i , j ) , where i and j are the number auxiliary gn and cs lines , respectively , linking the two fermion lines crossing the diagram . notice that there are not contribution from graphs of type ( 0,1 ) since they are not proper . figure [ fig6 ] furnishes examples of each one of these sets of diagrams . the final result for each order is obtained by summing the corresponding entries in each table . thus we have , @xmath139 s. weinberg , `` the quantum theory of fields '' , cambridge university press , 1995 ; g. p. lepage proceedings of tasi-89 , 1989 ; g. p. lepage _ et al . _ , d * 46 * , 4052 ( 1992 ) ; m. luke , a. v. manohar , phys . rev . d*55 * , 4129 ( 1997 ) . d. gross in methods in field theory , les houches 1975 , r. balian and j. zinn - justin eds . , north - holland , amsterdam , ( 1976 ) ; g. parisi , nucl . b100 , 368 ( 1975 ) ; b. rosenstein , b. j. warr and s.h . park , phys . lett 62 1433 , ( 1989 ) . m. gomes , r. s. mendes , r. f. ribeiro and a. j. da silva , phys . * d43 * , ( 1991 ) 3516 ; ; m. gomes , v. o. rivelles and a. j. da silva , phys . rev.*d41 * , 1363 ( 1990 ) ; n. v. krasnikov and a. b. kyatkin , mod . a6 , ( 1991 ) 1315 . b. rosenstein , b. j. warr and s.h . park , phys . d * 39 * 3088 , 1989 l. del debbio , s. j. hands , j. c. mehegan , nucl . b502 ( 1997 ) 269 ; s. hands , `` fixed point four - fermi theories '' , hep - lat/9706018 . we use natural units @xmath140 and our metric is @xmath141 . the fully antisymmetric tensor @xmath142 is normalized such that @xmath143 and we define @xmath144 . repeated greek indices sum from 0 to 2 , while repeated latin indices from the middle of the alphabet sum from 1 to 2 . for the @xmath9-matrices we adopt the representation @xmath145 , where @xmath146 are the pauli spin matrices .
in 2 + 1 dimensions , for low momenta , using dimensional renormalization we study the effect of a chern - simons field on the perturbative expansion of fermions self interacting through a gross neveu coupling . for the case of just one fermion field , we verify that the dimension of operators of canonical dimension lower than three decreases as a function of the chern - simons coupling .
the diffusion of dopants , such as arsenic , in silicon has been studied extensively because of its importance in integrated circuit manufacturing . native point defects , such as lattice vacancies , which may be created during implantation of the dopants , interact with other defects to facilitate diffusion of the implanted species . as circuit device sizes continue to shrink , approaching the scale of typical diffusion distances , a deeper understanding of this phenomenon is critical . the effects of temperature on the diffusion process have been studied experimentally and theoretically . the experimentally measured diffusivity of arsenic is known to increase as temperature increases.@xcite several theoretical models have , as expected , demonstrated the same result due to greater thermal fluctuations at higher temperatures . a previous kinetic lattice monte carlo ( klmc ) study@xcite showed how defect capture radii depend on interaction range and temperature . bunea and dunham@xcite used a klmc model with one arsenic , one vacancy , and interactions ranging from the third to sixth nearest lattice neighbor site . they found an increase in arsenic diffusivity at higher temperatures , and quantified deviations from a pair diffusion model in terms of a correction factor introduced into fick s first law@xcite . as temperature increased , the value of the correction factor deviated more from the expected value of 1 . they also demonstrated that deviations from pair diffusion increased with a longer interaction range . _ et al.__@xcite also used a klmc model with a single arsenic and vacancy and interactions extending from the third to 20th nearest neighbor to show that arsenic diffusivity increases with temperature . they demonstrated that a ring mechanism " , in which a vacancy moves along a hexagonal ring around an arsenic atom , contributes less to arsenic diffusion at higher temperatures and that this effect is more significant for longer interaction ranges . in section [ sec_temp ] we present the results of similar klmc calculations , with 1000 arsenic atoms and 100 vacancies and long range interactions . in section [ sec_range ] we study the effect of interaction range on diffusion and clustering . the effect of arsenic concentration on diffusivity and cluster formation has been studied because of its relation to the electrical deactivation of arsenic at high doping levels . fair and weber@xcite experimentally found maximum arsenic diffusivity with arsenic concentration of @xmath1 at @xmath2c . the diffusivity of arsenic decreased for concentrations above and below this level . _ et al.__@xcite found that arsenic diffusivity increased for arsenic concentrations above @xmath3 at @xmath4c . solmi and nobili@xcite demonstrated an increase in arsenic diffusivity for concentrations up to @xmath5 and constant diffusivity for higher concentrations . klmc simulations have also given differing results . dunham and wu@xcite found , using estimated third nearest neighbor interactions at @xmath4c , an increase in arsenic diffusivity at high doping levels , in agreement with larsen _ _ et al.__@xcite . bunea and dunham@xcite used a similar klmc model , with third nearest neighbor _ ab initio _ interactions at @xmath6c , and found enhancement over short simulation times , up to @xmath7 time steps , but the enhancement diminished over longer simulation times . list and ryssel@xcite also performed klmc simulations similar to those of dunham and wu@xcite but found no diffusivity enhancement . all of these klmc simulations used vacancy concentrations which depended on the temperature and the equilibrium vacancy concentration , @xmath8 . experiments@xcite performed at @xmath4c measured @xmath9 . the value most commonly used in klmc simulations , included those cited above , was @xmath10 , which is still a dilute concentration but large enough to allow for reasonable system sizes in klmc simulations . xie and chen@xcite summarized all these discrepancies , experimental and theoretical , and concluded that time dependent clustering effects were responsible . they performed _ ab initio _ calculations and showed that interactions to at least the ninth nearest neighbor were important and that a short term arsenic diffusivity enhancement would die out due to formation of clusters larger than arsenic - vacancy pairs . in section [ sec_conc ] , we present our klmc results , with long range ( 18th nearest neighbor ) _ ab initio _ interactions , for several arsenic concentrations , and a constant arsenic : vacancy ratio . in this work , we assume a large vacancy concentration , which would be expected post - implantation . in section [ sec_ratio ] we vary the arsenic : vacancy ratio to determine what effect vacancy concentration has on diffusion and clustering . we use a kinetic lattice monte carlo ( klmc ) method to simulate atomic scale diffusion processes . this method visits defects on a silicon lattice , either randomly or in some predetermined order , and , if a move is allowed , attempts to move each defect , one at a time . defects include substitutional dopants and native point defects , which include vacancies in this work . possible movements include vacancies exchanging lattice sites with lattice atoms or defects . although concerted exchange is a theoretical possibility@xcite , we do not consider it in this work . thus , a move is allowed if the defect in question is a vacancy , or if the defect in question is a dopant , and a randomly chosen neighboring site is occupied by a vacancy . if a move is attempted , the change in the system energy determines whether the move is accepted . if the move lowers energy , it is accepted . if the move does not lower energy , it is accepted , according to the metropolis detailed balance algorithm , with probability @xmath11 , where @xmath12 is boltzmann s constant , @xmath13 is the system temperature , and @xmath14 is the change in system energy as a result of the move . the system energy in each configuration is the sum of pair interactions between the defects . the values for the pair interactions at given separations on the lattice , out to 18 neighbors , were calculated with _ ab initio _ methods described in subsection b. the time scale of a klmc simulation is set by the hopping frequency , @xmath15 , of the defects , which is determined by @xmath16 where @xmath17 is the attempt frequency , @xmath18 is the height of the energy barrier for the move , @xmath12 is boltzmann s constant , and @xmath13 is the system temperature . in this work , for vacancies , we used @xmath19 calculated with the expression @xmath20 and the values @xmath21 , and @xmath22 , reported for low density vacancy diffusion by tang _ _ et al.__@xcite , and a lattice constant of 5.43 . since this value for @xmath18 is less than the 0.2 ev used by bunea and dunham@xcite , our vacancy hopping frequency is higher . for the attempt frequency of arsenic , we scaled the attempt frequency of vacancies @xmath23 where @xmath24 is the effective mass of a vacancy ( or the mass of a silicon atom ) and @xmath25 is the mass of an arsenic atom . we used 0.58 ev as the energy barrier for arsenic vacancy exchange , as reported by pankratov _ _ et al.__@xcite . after each klmc step , which visits all defects in the system , the simulation time is incremented by the time step of the fastest allowed event , which in this work is the exchange of a vacancy with a si lattice atom , set by @xmath26 . for events with longer time steps , such as the exchange of a vacancy with an arsenic atom , an attempt is made at each time step if a randomly generated number between 0 and 1 is less than , for example , @xmath27 klmc methods follow the position of each defect in the system over time . this allows the definition of a diffusion coefficient as @xmath28 which approaches the thermodynamic diffusion coefficient as @xmath29 approaches infinity . we use this definition in order to study diffusion during various time ranges . the pair interactions between vacancies@xcite and between vacancies and arsenic@xcite were calculated using the ultrasoft pseudopotential plane wave code vasp@xcite , on a 216 atom supercell . the generalized gradient approximation ( gga ) was used , along with a @xmath30 monkhorst - pack@xcite * k*-point sampling and a kinetic energy cutoff of 208 ev . all vacancies were charge neutral . the arsenic - arsenic interactions were calculated using petot@xcite , a norm conserving pseudopotential plane wave code , on a 64 atom supercell , using the local density approximation ( lda ) and a @xmath31 monkhorst - pack * k*-point sampling and a kinetic energy cutoff of 218 ev . in each case , the initial configuration of atoms was allowed to relax and the final energy was recorded for all possible separations of each defect pair . the results are shown in figure [ fig_potentials ] . [ 0.5 ] in this work , separate defects were not permitted to simultaneously occupy the same lattice site . for the following sections , we began with a baseline system containing 1000 arsenic atoms and 100 vacancies on a silicon lattice measuring 84 unit cells on each side for an arsenic concentration of @xmath0 . boundary conditions were periodic , and pair interactions to the 18th nearest neighbor were considered . the base system temperature was 900 k. in subsection a , we vary the temperature , while the interaction range varies in subsection b. in subsection c we consider various arsenic concentrations , and in subsection d we vary the arsenic : vacancy ratio . each simulation presented is the mean of five identical simulations . we define a cluster as two or more defects , substitutional arsenic or vacancies , occupying a nearest neighbor position from at least one other member of the cluster . [ 0.5 ] in this section we vary the system temperature while holding the number of arsenic and vacancies , the system size , and the interaction range constant . figure [ fig_d_t ] shows the diffusivity , calculated using equation ( [ eq_dt ] ) , of arsenic and vacancies as a function of time for temperatures ranging from 700 k to 1300 k. the quantity @xmath32 in equation ( [ eq_dt ] ) represents the continuing system time , and @xmath33 is reset to @xmath32 when @xmath34 is twice as large as the previous time interval . we see that the vacancies diffuse freely at first , while the arsenic atoms are stationary . at all temperatures the free vacancy diffusivities match the theoretical values , which are shown on figure [ fig_d_t ] for times earlier than the first data point on the vacancy diffusivity curves . as arsenic - vacancy pairs ( asv ) form , the arsenic begins to diffuse . the diffusivity of both species is greater at higher temperatures , as expected , since the thermal fluctuations @xmath35 are larger relative to the migration energy of both asv and free vacancies , and by @xmath36 seconds the transient effects disappear . [ 0.5 ] [ 0.5 ] in figure [ fig_free_t ] , the fraction of the total number of defects which are free ( not in a cluster ) is shown over time for each species . at higher temperatures , fewer defects are in clusters . the same trends can be seen in figure [ fig_ndefs_t ] , which shows the number of clusters and the number of defects per cluster over time . more clusters form at lower temperatures and these are largely arsenic - vacancy pairs ( asv ) for times up to approximately @xmath37 seconds . at that time the number of clusters peaks for lower temperatures , and the number of defects per cluster rises above two for all temperatures . at higher temperatures , fewer clusters form , but those that do form are slightly larger on average . [ 0.5 ] figure [ fig_fractv_t ] shows that the vacancy fraction of defects in clusters increases over time for all temperatures considered . for lower temperatures , this implies that asv begin to aggregate , as indicated in figure [ fig_ndefs_t]a , and that some of the asv dissociate , as in figure [ fig_free_t]a , with the vacancy attaching itself to another asv . at higher temperatures the number of clusters is more stable over time , as in figure [ fig_ndefs_t]a , which implies that mobile asv attract free vacancies . in either case , a pair diffusion model is only valid for a short time , roughly @xmath37 seconds . this result expands the work of bunea and dunham@xcite , who found little deviation from the pair diffusion model at lower temperature but did not consider time dependence . it should be noted that the results of bunea and dunham were obtained by simulating a single arsenic - vacancy pair with interactions ranging from third to sixth nearest lattice neighbor . in order to study the effect of interaction range on arsenic diffusion and clustering , we truncated and shifted the interaction potentials in figure [ fig_potentials ] from 18 shells to 12 , 6 , and 4 shells , so that the interaction vanishes at the indicated separation . the system temperature , the number of arsenic and vacancies , and the system size remained constant . [ 0.5 ] figure [ fig_d_range ] shows the diffusivity of arsenic and vacancies , again calculated using equation ( [ eq_dt ] ) , over time for interaction ranges of 18 , 12 , 6 , and 4 shells . those simulations with the longest range interactions show an initial transient behavior until @xmath36 seconds , after which the diffusivity is constant over time . in contrast , for shorter interaction ranges , the diffusivity continues to decrease over time . [ 0.5 ] figure [ fig_free_range ] shows the fraction of each species free over time . for all interaction ranges considered , the fraction of free arsenic decreases as arsenic - vacancy pairs ( asv ) form , but for 4 , 6 , and 12 shell interactions , almost all arsenic are free from clusters at later times and thus immobile . for 18 shell interactions , the fraction of arsenic free remains roughly constant over the same time range that arsenic diffusivity is constant . for vacancies , 18 shell interactions allow some vacancies to remain free at all times , while shorter range interactions lead to a state in which all vacancies in the system are bound in clusters . [ 0.5 ] figure [ fig_ndefs_range ] presents the number of clusters and defects per cluster over time for the interaction ranges of interest . when considering 18 shell interactions , both quantities reach a somewhat constant level over time , while all shorter range interactions cause the number of clusters to continue to decrease and the size of the clusters to increase . combined with the free fractions in figure [ fig_free_range ] , this implies that asv form initially , for times up to @xmath37 seconds , after which , as asv begin to interact , the short range vacancy - vacancy interaction overwhelms all other interactions ; the asv dissociate , the arsenic atoms are left isolated and immobile , and the vacancies cluster with each other . [ 0.5 ] this can be seen more clearly in figure [ fig_fractv_range ] ; for shorter interaction ranges , almost all defects in clusters are vacancies at long times . the strong interaction between vacancies at short range dominates the clustering behavior of the system if no longer - range interactions are included . [ 0.5 ] the concentration of dopants in the system can be varied by two methods . either the dimensions of a simulation cell with periodic boundary conditions or the number of dopants in the cell can change . the diffusivities for both arsenic and vacancies , calculated using equation ( [ eq_dt ] ) , are virtually identical over time for each method at high concentrations . at low concentrations , some time dependent differences emerge in diffuse systems . we chose to vary arsenic concentration by changing the size of the simulation cell , with periodic boundary conditions , maintaining a constant number of arsenic and vacancies in the system . this method is much more computationally efficient , as the simulation time increases with the number of defects in the system . the system temperature and the interaction range are constant , as well . figure [ fig_d_c ] shows the diffusivity , calculated using equation ( [ eq_dt ] ) , of arsenic and vacancies over time for arsenic concentrations ranging from @xmath38 to @xmath39 . the arsenic diffusion begins as arsenic - vacancy pairs ( asv ) form , which happens earlier as the arsenic concentration increases . at the highest arsenic concentration considered , @xmath39 , there is almost no free diffusion of vacancies , due to the long range interactions . figure [ fig_d_c]a shows that , while arsenic diffusivity is initially higher at greater arsenic concentrations , by @xmath36 seconds , the maximum arsenic diffusivity occurs for concentrations in the range of @xmath40 to @xmath0 , while concentrations greater and less than this range yield a lower diffusivity . this is the same trend observed experimentally by fair and weber@xcite , although the peak diffusivity measured in that work was at @xmath41 as @xmath42 , at a temperature of @xmath2c . the results in figure [ fig_d_c]a show a different trend than the experimental work of larsen _ _ et al.__@xcite and the klmc results of dunham and wu@xcite . both of those studies , conducted at @xmath4c , found an increase in arsenic diffusivity for dopant concentrations above @xmath43 . the simulations of dunham and wu considered third nearest neighbor interactions which decreased linearly with distance . our results also disagree with those of bunea and dunham@xcite , who found greater enhancement in arsenic diffusivity at higher arsenic concentrations for short simulation times at @xmath6c ; the effect they reported decreased over long simulation times . it should be noted that the simulations of bunea and dunham considered interactions ranging from the third to sixth nearest lattice neighbor , in contrast to our long range interactions . [ 0.5 ] figure [ fig_nclusters_c ] shows that the number of clusters over time is greater as arsenic concentration increases , and that the typical cluster size is roughly the same for different doping levels . this alone does not explain why arsenic diffusivity is not enhanced at greater concentrations , but figure [ fig_fractv_c ] shows that as the arsenic concentration increases , the fraction of vacancies in clusters decreases . thus , at higher doping levels the arsenic fraction of defects in clusters increases . this supports the conclusions of xie and chen@xcite and bunea and dunham@xcite who suggested the development of large , arsenic dominated clusters at elevated doping levels . [ 0.5 ] immediately after ion implantation , the concentration of vacancies should be much higher than the equilibrium concentration . we have assumed a baseline vacancy concentration of @xmath40 in this work , for an arsenic : vacancy ratio of 10:1 . in order to explore the ramifications of this choice , we performed klmc simulations at several different arsenic : vacancy ratios , ranging from 2:1 to 200:1 , with a constant arsenic concentration of @xmath0 . the arsenic : vacancy ratio of 200:1 corresponds to the equilibrium vacancy concentration used in previously cited klmc works . the system temperature , the interaction range , and the size of the simulation cell were held constant . [ 0.5 ] figure [ fig_d_ratio ] shows the diffusivities of arsenic and vacancies , calculated using equation ( [ eq_dt ] ) , for various arsenic : vacancy ratios . the diffusivity of arsenic is greater with more vacancies in the system . conversely , vacancy diffusivity is lower with more vacancies in the system . [ 0.5 ] figure [ fig_free_ratio ] supports this trend . with fewer vacancies in the system , a greater fraction of both arsenic and vacancies will be free over time . [ 0.5 ] [ 0.5 ] in figure [ fig_ndefs_ratio ] we see that a greater number of vacancies produces both a higher number of clusters and larger clusters , and figure [ fig_fractv_ratio ] shows that the vacancy fraction of defects in clusters will increase with vacancy concentration . thus , having more vacancies in the system leads to more clusters , larger clusters , and more vacancies in the clusters . this is most likely due to the strong vacancy - vacancy interaction at short range . we have presented results of kinetic lattice monte carlo simulations using long range interactions , to the 18th nearest lattice neighbor , and large numbers of defects , including 1000 arsenic atoms and a high vacancy concentration , which would be found shortly after implantation of dopant atoms into silicon . we demonstrated that the trend at higher temperatures is toward fewer clusters with more defects . at all investigated temperatures , the vacancy fraction of defects in clusters increases over time . at low temperatures , this implies the dissociation of some arsenic - vacancy pairs ( asv ) , after which the newly free vacancy attaches to a nearby asv . at higher temperatures , mobile asv attract free vacancies . in both cases , a diffusion model based only on arsenic - vacancy pairs may be inadequate because of time dependent clustering effects . we also demonstrated that considering only short interaction ranges , less than 18 shells , produces mostly vacancy clusters , due to the strong short range attraction between vacancies . this behavior manifests itself in decreasing diffusivities over time , exclusion of arsenic from clusters as asv dissociate , decreasing number of clusters while cluster size grows , and vacancy fraction of defects in clusters approaching 1 over time . such behavior implies that little or no arsenic diffusion is occurring and is in contradiction with experimental evidence . therefore , long range interactions seem crucial for realistic simulations of defect migration in silicon . we showed that arsenic diffusivity increases for arsenic concentrations up to @xmath0 and then decreases as arsenic concentration increases . this behavior is due to the formation of a greater number of clusters at higher arsenic concentration ; the size of the clusters does not increase substantially . the vacancy fraction , however , changes significantly , transitioning from vacancy dominated clusters to arsenic dominated clusters at high arsenic concentrations , keeping the arsenic : vacancy ratio constant . lastly we studied diffusion and clustering behavior as a function of the number of vacancies in the system . we concluded that higher concentrations of vacancies lead to larger clusters , more clusters , and more vacancies per cluster . while a large vacancy concentration is expected immediately after ion implantation , this may not be realistic at later times . more work is needed to incorporate interstitial defects , especially silicon self interstitials so that frenkel pair recombination can change the vacancy concentration over time . we thank keith m. beardmore for contributing the vasp interaction potentials , lin - wang wang and andrew canning of the lawrence berkeley national laboratory for the petot code and advice on the calculations , and mike masquelier and wolfgang windl for useful discussions in the early stages of the work . initial part of the work was supported by the computational nanoscience group of motorola , inc .
we present results of kinetic lattice monte carlo ( klmc ) simulations of substitutional arsenic diffusion in silicon mediated by lattice vacancies . large systems are considered , with 1000 dopant atoms and long range _ ab initio _ interactions , to the 18th nearest lattice neighbor , and the diffusivity of each defect species over time is calculated . the concentration of vacancies is greater than equilibrium concentrations in order to simulate conditions shortly after ion implantation . a previously unreported time dependence in the applicability of the pair diffusion model , even at low temperatures , is demonstrated . additionally , long range interactions are shown to be of critical importance in klmc simulations ; when shorter interaction ranges are considered only clusters composed entirely of vacancies form . an increase in arsenic diffusivity for arsenic concentrations up to @xmath0 is observed , along with a decrease in arsenic diffusivity for higher arsenic concentrations , due to the formation of arsenic dominated clusters . finally , the effect of vacancy concentration on diffusivity and clustering is studied , and increasing vacancy concentration is found to lead to a greater number of clusters , more defects per cluster , and a greater vacancy fraction within the clusters .
the manganites lnbamn@xmath7o@xmath8 present a novel class of layered materilas , in which the layered arrangements of y and ba cations results into the regular architectures of the half - occupied e@xmath9-orbitals of mn@xmath10 with the out - of - plane and in - plane orientations for @xmath11 and @xmath12 , respectively . both series enclose manganese in state of half - doping mixed valence . the orbital order is coupled to charge order , consisting in the alternations mn@xmath10/mn@xmath13 and mn@xmath10/mn@xmath14 for @xmath11 and @xmath12 , respectively , along three axes , which are cartesian for the majority of ln s , but slightly oblique in monoclinic@xcite or triclinic@xcite ybamn@xmath7o@xmath3 . owing to the layered structure the polycrystalline materials are highly susceptible to preferred orientation withstanding a simple quantitative description , except the special case of ybamn@xmath7o@xmath3 . recently , such a quantitative description was suggested@xcite to be useful in vibrational spectroscopy on powders of anisotropic materials . as a matter of fact , the oriented polycrystals consisting of platy crystallites can replace the single crystals in studies of a variety of anisotropic properties of materials , such as electric , magnetic or vibrational . especially , when the polycrystalline material is a ferromagnet or a superconductor it can be subjected to a thorough texturing in an external magnetic field . [ ptb ] f1.eps in this work , members of both the `` o@xmath2''@xmath15@xcite and `` o@xmath3''@xmath15 @xcite families with ln = y , gd , sm , nd , pr , la and ln = sm@xmath16nd@xmath17 solid solutions were doped with 2% of @xmath18fe and investigated using rietveld analysis , magnetization and mssbauer spectroscopy on both random and oriented powders . we investigate the effect of the fe substituents on the phase transitions known to occur in undoped systems . clean samples exhibit the coo on the long - range scale observed previously via neutron diffraction@xcite . we show that the substitution smears some of the transitions and lowers the temperature for others . in charge - ordered state , the fe substituents display only one component fe@xmath10 in mssbauer spectra . with doping the stable - valence ion fe@xmath10 into the mixed - valence site of jahn - teller ions , such as mn@xmath19/mn@xmath10or mn@xmath20mn@xmath14 , the long - range coo becomes the subject of suppression by quenched disorder . the short range order is understood to preserve a favorite arrangements of mn electronic configurations around the @xmath18fe impurity dopants . via the quadrupole interactions and isomer shifts we explore the variety of structural and valence states adopted by the impurity . in addition , due to the layered structure the manifestations of vibrational anisotropy appears in mssbauer line intensities , similarly to layered cuprates , which showed a notable goldanskii - karyagin effect ( gke)@xcite . in samples with preferred orientation , the texture effects are combined in mssbauer spectra intensities with gke and a technique is suggested to disentangle these effects . the same technique can be useful to determine the anisotropy in tensor material properties starting from the data obtained on the aligned powders . we illustrate the application of this technique to one of such properties , when the measurements is feasible again with the radiation of the same wavelength as in mssbauer spectroscopy , however , realizable only on a large - scale synchrotron facility . this property is the anisotropic multicomponent phonon density of states ( dos ) , which request the single crystals for ordinary measurement@xcite , however , this work presents a proposal for determination of both dos components using the aligned powders instead of a single crystal . the layered a site - ordered oxygen - saturated manganites lnbamn@xmath21fe@xmath1o@xmath3 were prepared for ln = y , gd , sm , ( sm@xmath22nd@xmath23 ) , ( sm@xmath23nd@xmath22 ) , nd , pr and la . the oxides ln@xmath7o@xmath24 and fe@xmath25o@xmath24 were mixed with carbonates baco@xmath24 and mnco@xmath24 . these mixtures were first annealed in 6n pure ar ( 99.9999% ) flow at 1350@xmath26c . this annealing has led to obtaining the oxygen - depleted phases lnbamn@xmath21fe@xmath1o@xmath2 . the synthesis of the oxygen - saturated `` o@xmath3 '' manganite is a two - step process as described previously@xcite . the second step consisted of annealing in o@xmath7 flow at low temperature ( 500@xmath26c ) . while in the first step of the oxygen - depletion the layered structure is formed of the alternating bao and y sheets interleaved by the mno@xmath7 checkerboards , in the second step this structure is intercalated into y - layer with the additional oxygen . a third family of manganites with the same cationic compositions was obtained in air ( without layer - growing treatment in 6n ar ) and showed a simple perovskite - like disordered structure . [ ptb ] f2.eps rietveld profiles for both `` o@xmath2 '' and `` o@xmath3 '' families of layer - ordered manganites were obtained by means of a `` mac science '' diffractometer using cu-@xmath27 radiation ( @xmath28= 0.15405 nm and 0.15443 nm ) . mssbauer spectra were measured at room temperature . isomer shifts are referred relatively @xmath29fe . to prepare the oriented powder samples several methods were used . the powders were wet - spread in alcohol and dry - spread onto scotch tape or onto blotting - paper . the data were obtained for powders oriented with their largest dimension in the plane - of the tape with a marked texture . rietveld refinements for all diffraction profiles were carried out using fullprof program@xcite . measurements of magnetization were performed using a squid magnetometer in an applied field of 1 koe at heating the samples from 5 k to @xmath30 and then at cooling from @xmath31 down to 5 k. this measurment protocol was applied in lnbamn@xmath0fe@xmath1o@xmath3 for ln = sm with @xmath32=400 k and for ln=(nd@xmath22sm@xmath23 ) with @xmath33370 k. the magnetization in ybamn@xmath0fe@xmath1o@xmath3 was measured first at heating from ambient temperature to @xmath32= 600 k ( ln = y ) and then at cooling from @xmath31 down to 5 k. the sample was then remagnetized at 5k by setting the external field @xmath34 koe followed by reapplying @xmath35 koe . finally , this sample magnetization was measured at heating up to 300 k. prior to first reports on the manganites ybamn@xmath7o@xmath2@xcite and labamn@xmath7o@xmath8 ( @xmath11 and @xmath36)@xcite which crystallized in the perovskite - like bilayered tetragonal structures , the ferrocuprate ybacufeo@xmath2 with the same structure built up of bipyramidal layers was already known@xcite since 1988 . on the other hand , the oxygen - rich ( @xmath12 ) layered manganite system had no prototype among cuprates . there exist two important features of the manganite systems , distinguishing them from the cuprate one . first , the layers of ba and ln develop in labamn@xmath7o@xmath8 only if the system is deoxygenated ( @xmath37 ) at high - temperature synthesis . excess of oxygen during the high - temperature annealing breaks the layered structure and create the disordered isotropic phase . second , both the `` o@xmath2 '' and `` o@xmath3 '' structures exhibit the in - plane ordering mixed valence ions mn@xmath13/mn@xmath38 and mn@xmath39/mn@xmath14 , respectively , both associated with coo . we observed that all of our x - ray profiles from `` o@xmath2 '' samples are perfectly fitted with the @xmath40 model ( fig.1 ) . the quality of fit was strongly declined when we attempted to fit the patterns in terms of @xmath41 model . this result suggests that the coo is essentially preserved in our samples having the fe ions placed into 2% of the mn sites . however , from mssbauer single - site spectra a coo meltdown is suggested , as shown below . such a meltdown is not denotiative of full randomness . short range order in the arrangement of charges and orbitals is preserved . indeed , since no superstructure reflections were observed , associated with lowering symmetry from @xmath41 to @xmath40 , our rietveld refinement preference for the symmetry group @xmath40 indicates only that the coo is kept on a short range . in pure labamn@xmath7o@xmath42 the supercell reflections are observable with electron diffraction@xcite , while fe - doping may broaden them and suppress their intensity to the level insufficient for observation . the coexistence of the preserved short - range coo with the doping - induced quenched disorder is discussed below along with mssbauer spectra . [ ptb ] f3.eps in contrast to the singular `` o@xmath2''-phase ( @xmath43phase@xmath44 , that is common for all the rare - earth elements , a few different modifications are known for the oxygen - saturated `` o@xmath3 '' a - site ordered layered manganites , depending on the size of ln @xcite . all of them are built up of octahedral bilayers composed of octahedra distorted in one way or other . our rietveld refinements were carried out using several symmetry groups and structure models known from the literature@xcite . these treatments have shown that the tetragonal cell @xmath45 @xmath46 @xmath47 @xmath48suits best ( fig.2 ) to all the samples in the layer - ordered oxygen - saturated family , with one exception of ybamn@xmath0fe@xmath1o@xmath3 . the structure of the latter was refined using the monoclinic symmetry ( space group p2 , no.3 ) . [ ptb ] f4.eps to smmarize , the lattice cells were fitted with the volume double or quadruple of perovskite one ( @xmath49 or 4@xmath50 ) owing to the lattice parameters relations : [ c]l@xmath51 + @xmath52 + @xmath53 for @xmath54 comparing the parameters of reduced lattice cell ( i.e. a distorted perovskite cell ) one can analyze the extent of distortion depending on ln , in function of the volume of the reduced cell . fig . 3 summarizes the results of refinement of the lattice cell parameters . the reduced parameters @xmath55 ( @xmath56 ) and @xmath57 are plotted against the reduced cell volume @xmath58 ( @xmath59 ) . from y to la the volume @xmath60 varies in lnbamn@xmath0fe@xmath1o@xmath2 within 5% and both @xmath61 and @xmath62 vary in the range of 1.7% . distortion of the reduced cell can be calculated in lnbamn@xmath7o@xmath2 as @xmath63 . from y to la the distortion decreases from 2.56% to 2.36% , respectively . in the oxygen - saturated series , the distortion @xmath64 does not show a stationary level from y to la , but changes abruptly between sm and nd . we observe in fig . 4 that there occurs in lnbamn@xmath0fe@xmath1o@xmath3 the phase transitions associated with coo and magnetic ordering . the same jumps and humps of susceptibility are observed as in the undoped lnbamn@xmath25o@xmath3 , however , the corresponding transition temperatures are suppressed . according to size of ln three groops of the oxygen - saturated manganites were specified previously@xcite as ln(i ) @xmath65(y , tb , dy , ho ) , ln(ii)=(sm , eu , gd ) and ln(iii)=(nd , pr , la ) . a member from each of these families was investigated in this work for the effect of the fe substituents on the transitions manifested in magnetic properties . the transition sequences ( temperatures @xmath66 , @xmath67 ) in the lightly ( i.e. 2% ) fe - doped manganites are essentially same as in undoped manganites , but differ in details . in ybamn@xmath0fe@xmath1o@xmath3 and smbamn@xmath0fe@xmath1o@xmath3 , the highest in temperature jump of magnetzation is associated with a structural transition , which is monoclinic - to - triclinic@xcite or triclinic - to - monclinic@xcite for ln = y , and tetragonal - to - orthorhombic@xcite for ln = sm . orbital ordering is now believed@xcite to accompany these structural transition ( @xmath68 ) , while complete charge ordering is attributed@xcite , e.g. for ln = y , to a separate small hump shifted from @xmath66 to lower temperature by @xmath69 k. in the undoped manganites , the temperature @xmath70 is associated with sharp localization of charge carriers . in the magnetization of ybamn@xmath21fe@xmath1o@xmath3 , the large jump is observed apparently without any small foregoing hump . interestingly , similar disappearance of the hump in magnetization caused by fe doping was reported for orbital ordering transition in bimno@xmath24@xcite . smearing the transition over a broad temperature range induced by fe substitution suppresses such a hump . another key feature of the doped systems stems from the fact that the values of @xmath66 are slightly suppressed compared to undoped ybamn@xmath25o@xmath3 and smbamn@xmath7o@xmath3@xcite . the suppression ranges are @xmath712%fe@xmath7250k and 40 k for ln = y and sm , respectively . in contrast to similar values of @xmath73 , there occurs a large difference between the cases of ln = y and sm for the shift of the transition temperature with the reversal of temperature sweep direction . such a shift is associated with an energy barrier for nucleation of a new phase within the region of overheating or undercooling the preceding phase . the hysteresis indicates strongly the first - order character of transition that was observed also in undoped ybamn@xmath7o@xmath3@xcite . the large nucleation barrier is observed in ybamn@xmath7o@xmath3 but not in smbamn@xmath7o@xmath3 . this is in agreement with a very small structural distortion in smbamn@xmath7o@xmath3 at @xmath66 as reported by akahoshi et al@xcite . temperatures of neel ( @xmath67 ) are also notably suppressed in both cases , as well as in the third group member , nd@xmath22sm@xmath23bamn@xmath74fe@xmath1o@xmath3 . the antiferromagnetic transitions humps were observed@xcite at 290 k and 250 k in undoped ndbamn@xmath7o@xmath3 and smbamn@xmath7o@xmath3 , respectively , therefore , the @xmath67 value of 286 k is expected for the solid solution nd@xmath22sm@xmath23bamn@xmath7o@xmath3 . remaining suppression @xmath75 @xmath76 k should be attributed to the effect of fe substitution . temperature ranges for sweep - reversal hysteresis around @xmath67 are not much different between the three . in both series of the layered manganites the samples displayed some degree of preferred orientation which is directly seen in x - ray diffraction patterns through an enhancement of the reflections of the type @xmath77 ( fig.2 ) . this indicates the platy habits of thin crystallites with the easy cleavage planes parallel to the layers . table 1 summarizes the results of refinement of the preferred orientation parameters . * table 1 . * parameters of march - dollase function , @xmath78 and @xmath79 , eq.(1 ) , refined from x - ray diffraction patterns . [ c]l|llllll| & & + & g@xmath80 & & & & & + & 0.61(3 ) & 0.27(2 ) & 5.64 & n / a * * * & n / a * * * & + & 0.61(1 ) & 0.47(2 ) & 3.88 & 0.57(3 ) & 0.84(2 ) & 1.88 + & 0.73(1 ) & 0.68(4 ) & 1.71 & 0.76(3 ) & 0.78(2 ) & 1.38 + & 0.61(2 ) & 0.74(3 ) & 2.18 & 0.74(3 ) & 0.76(2 ) & 1.48 + & 0.80(2 ) & 0.59(10 ) & 1.57 & 0.97(5 ) & 0.99(1 ) & 1.00 + & 0.47(4 ) & 0.75(5 ) & 3.80 & 0.55(10 ) & 0.94(3 ) & 1.35 + & 0.45(6 ) & 0.57(10 ) & 7.56 & 0.46(22 ) & 0.97(2 ) & 1.31 + & 0.68(2 ) & 0.75(15 ) & 1.74 & 0.96(3 ) & 0.86(3 ) & 1.03 + * 1 : sm@xmath22nd@xmath23bamn@xmath0fe@xmath1o@xmath81 * * 2 : sm@xmath23nd@xmath22bamn@xmath0fe@xmath1o@xmath17 * * * refinement with the uniaxial texture along [ @xmath77 ] was not applicable by reason of monoclinic structure . the march - dollase ( md ) function@xcite depends on three variables : the texture - axis - misfit angle @xmath83 and two profile - refinable parameters @xmath78 , @xmath79 ( ref . ) , @xmath84 ^{\frac{3}{2 } } } \label{1}%\ ] ] that is to say , @xmath83 is the acute angle between the x - ray scattering vector @xmath85 and the axis of preferred orientation . in powders of single - crystalline particles , the preferred axis is dictated by the grain shape . the cylindrical - symmetry texture axis runs along the whisker - like or fiber - like crystals , but along the _ normal _ to the plate - like crystals . we fitted all our x - ray patterns with the uniaxial texture along the direction [ @xmath77 ] . the eq.([1 ] ) describes the density of poles , which come into reflection position at a given bragg angle @xmath86 . from the viewpoint of conserving the scattering matter the function @xmath87 is a true distribution function suitable for quantitative analysis . in this sense , the march - dollase distribution is much better than gaussian distribution , originally suggested by rietveld@xcite . the parameter @xmath79 desrcibes the fraction of the sample that is not textured and the parameter @xmath78 describes the degree of alignment within the textured fraction . the diffraction intensities are scaled by the distribution @xmath87 which culminates at @xmath88 for @xmath89 , but at @xmath90 for @xmath91 . therefore , both plate - like and needle - like powders are refinable with the universal function @xmath87 . the domains @xmath92 and @xmath93 correspond to the platy and acicular habits , respectively , while fully isotropic powder fits to @xmath94 . two parameters of the textured sample anisotropy @xmath78 and @xmath79 presents the full set of texturing characteristics , however , it is convenient to simplify the sample comparison using the single parameter @xmath95 given by a combination of @xmath78 and @xmath79 . this ratio displays directly the enhancement of right - hand line within the closely - spaced pairs @xmath96 and @xmath97 ( fig.2 ) . from the table 1 it becomes clear that lnbamn@xmath0fe@xmath1o@xmath2 shows much stronger tendency to texturing than lnbamn@xmath0fe@xmath1o@xmath3 . this result could be explained by variations in crystallite size and aspect ratio . [ ptb ] f5.eps returning to the preparation conditions of these series of samples , one sees that lnbamn@xmath0fe@xmath1o@xmath3 was prepared from lnbamn@xmath21fe@xmath1o@xmath2 in oxygen at the temperature as low as 500@xmath26c . in such a low temperature process , the crystallites could be destroyed and diminished in size , but hardly could grow . it follows indeed from fig.3 that the large compression of the lattice takes place at oxygenation . it was observed previously in gdbamn@xmath7o@xmath98 @xcite , that the rate of oxygen intercalation into the layered oxide is extremely high . strains associated with the oxygenation are very large . the crystal cracking and breakup in lateral dimension are likely to accompany the oxygen intercalation . the crystal cracks and strains observed during oxygenation were attributed to orthorhombicity@xcite . we have stabilized the orthorhombic phase gdbamn@xmath7o@xmath99 using a more complicated sequence of thermal treatments@xcite . the orthorhombic distortion in gdbamn@xmath7o@xmath99 is indeed much larger than either d@xmath2 or d@xmath3 , calculated above . large distortion arises from channel - like structure of the half - filled layer lno@xmath100 and this could be the plausible origin of the crystal dispergating with oxygenation@xcite . mssbauer spectra in oriented samples of lnbamn@xmath0fe@xmath1o@xmath2 are clearly asymmetric . in the samples with ln = gd and ln=(sm@xmath22nd@xmath101 ) the spectra are fitted with single asymmetric doublet . however , two doublets are crucial to fit the spectra for ln = la and sm ( fig.5 ) . spectral asymmetry increases from la towards gd and this increase correlates with increasing distortion d@xmath2 . the second doublet is not enough resolved to fit separately its asymmetry . therefore , its asymmetry was fixed at fitting to be the same as the asymmetry of the major doublet . as the size of ln decreases we observe the systematic decrease in the area of second doublet . an exception from this series is made by the mixed rare - earth manganite sm@xmath22nd@xmath23bamn@xmath0fe@xmath1o@xmath2 showing a single doublet , although the average size of ( sm@xmath22nd@xmath23 ) is larger than the size of sm , while smbamn@xmath0fe@xmath1o@xmath2 still shows a presence of the second doublet . however , this observation is in line with our interpretation of the single - site spectrum as originating from the meltdown of the charge order induced by the pointlike quenched disorder . indeed , the additional pointlike disorder in sm@xmath22nd@xmath23bamn@xmath0fe@xmath102o@xmath2 is related to disorder in the ln site . [ ptb ] f6.eps an important result appears in the fact that fe sees the charge order broken more easily for smaller ln by the same level of doping . the double - site spectrum in labamn@xmath0fe@xmath1o@xmath2 evidences that the charge order is robust for ln = la , so that the fe species ocuppy both mn(ii ) and mn ( iii ) sites , although the area ratio is 7:3 instead of 5:5 expected for random occupation . such a random occupation might be expected only in case of long - range charge order . however , iron can adopt more easily its preferred site when the charge and orbital ordered domains decrease in size . there occurs enough charge fluidity in gdbamn@xmath0fe@xmath1o@xmath2 and in sm@xmath22nd@xmath23bamn@xmath0fe@xmath1o@xmath2 to realize the unique surrounding for overwhelming majority of dopants . on the other hand , in labamn@xmath0fe@xmath1o@xmath2 and in smbamn@xmath0fe@xmath1o@xmath2 , we observe the robust coo that has a longer correlation lengths , such that a minor but significant part of dopants remains in a different surrounding . the second doublet shows the same isomer shift but twice increased quadrupolar splitting . both doublets originate from @xmath18fe in pyramidal coordination , however , one of these doublets originates from the site having a larger apical distance and stronger jahn - teller character . trying to understand the origin of difference between doping behaviors for large and small ln s , let us compare the pyramidal interatomic distances in labamn@xmath7o@xmath2 and in ybamn@xmath7o@xmath2 as reported previously@xcite . the difference between mn(ii ) and mn(iii ) pyramids is similar for ln = la and y , however , in ybamn@xmath7o@xmath2 , both mn(ii ) and mn(iii ) pyramids are more elongated in apical dimension , and compressed in equatorial dimension . such a distortion appears thus to facilitate the melting of the charge and orbital order around the doped fe species . [ ptb ] f7.eps known origins of the line area asymmetry in mssbauer spectra are the preferred orientation of crystallites and vibrational anisotropy . in a random polycrystalline material , the ordinary single - site spectra in paramagnetic domain are the symmetric doublets . the equal intensities of both lines result from random - powder averaging , unless gke takes place . preferred orientation would induce the asymmetry with or without gke . we discuss first the separate ocurrence of texture effects and gke and start from the texture effects . [ [ texture - effects ] ] texture effects + + + + + + + + + + + + + + + let us consider the doublet intensity ratio for a single crystallite . this ratio depends on the orientation of the wave vector of the incident x - ray quantum with respect to the axes of the electric field gradient ( efg ) tensor of the site wherein the @xmath18fe nucleus is located . the angular dependence of clebsch - gordan coefficients@xcite results in the @xmath83-dependent intensity ratio of the quadrupole doublet @xmath103 for each crystallite : @xmath104 when the unoriented powder is used for measuring the spectra the angular averaging gives @xmath105 , @xmath106 and @xmath107 . in case of oriented powder , using the md texture function ( eq.[1 ] ) , we substitute @xmath108 and @xmath109 @xmath110 with @xmath111 where@xmath112 both factors @xmath113 and @xmath114 are imaginary for @xmath92 , however , @xmath115 is real in full range @xmath116 . rietveld analysis allows one to fit the parameters of the degree of alignment @xmath78 and the aligned fraction @xmath79 starting from a set of x - ray diffraction intensities . a similar problem can be formulated in mossbauer spectroscopy : to determine @xmath78 and @xmath79 starting from the spectral asymmetry . in the area of small asymmetries , the effect of @xmath78 and @xmath79 on the asymmetry turns out to be nearly equal , as will be shown below . in fig.6 , the degree of alignment @xmath78 is plotted vs. the asymmetry @xmath103 for several values of the textured fraction @xmath79 . [ [ vibrational - anisotropy ] ] vibrational anisotropy + + + + + + + + + + + + + + + + + + + + + + the vibrational anisotropy @xmath117 is also plotted in fig.6 vs. @xmath103 . it is expressed through the absolute value of the wave vector for @xmath118-radiation and the mean - square vibrational displacements along @xmath119 , and in perpendicular direction , @xmath120 . this plot was calculated following the integral of each line intensity scaled with the angle - dependent lamb - mssbauer factor for uniaxial symmetry @xmath121 : @xmath122 the eq.([22 ] ) is appropriate in randomized powders of all our `` o@xmath3 '' phases , except ybamn@xmath0fe@xmath1o@xmath3 . the sigmoidal function @xmath123 is decreasing from 3 to 0.6 when @xmath117 is varied from @xmath124 to @xmath125 . the inverse function @xmath126 is shown in fig.6 as a semilog plot , having two branches . the right - hand branch ( @xmath127 ) corresponds to oblate , pancake - shaped ellipsoids of atomic thermal displacements ( atd ) , and the less - dispersive left - hand branch correspond to prolate , or cigar - like atd . from fig.6 it becomes clear that the underbalanced line area ratio ( @xmath128 is related to either prolate atd or to acicular crystallite textures . the overbalanced ratio ( @xmath129 ) is produced either by oblate atd or by platy crystallite textures . mixing these effects will be analysed below in 8 . [ ptb ] f8.eps [ ptbptb ] f9.eps the case @xmath129 refers in eqs.([033],[22 ] ) to stronger line @xmath130 and weaker line @xmath131 . the line @xmath130 lies higher in energy than the line @xmath132 if @xmath5 . _ vice versa , _ if @xmath6 , the lines swap the energy positions . we attribute the more intense line in fig . 5 to the transition @xmath133 assuming that the electric field gradient @xmath5 is oriented along @xmath134axis perpendicular to layers . this is in agreement with the point charge model for pyramidal coordination @xcite . thus , either oblate atd or platy crystallite textures underlie the observed in fig.5 spectral asymmetry . among the reports of neutron diffraction profile refinements in lnbamn@xmath7o@xmath2 , no data are known for an anisotropic atd because all profiles were refined using b@xmath135@xcite . the anisotropic atd , however , were refined previously for pyramidal coordination of ybafecuo@xmath2 . the oblate atd ellipsoids were never observed in pyramidal coordination . on the opposite , the atd tensor shaped as very extended prolate cigar was found in ybafecuo@xmath2@xcite . thus , the self - consistent combination is @xmath5 and prevailing asymmetry owing to platy - crystallite texture . the asymmetry would be reduced by prolate atd , but not fully compensated . the compensation effects can be described quantitatively if we find what texture parameters are equivalent to vibrational anisotropy parameter in its effect on asymmetry . having arrived to the quantitative description of texture using the eqs.([06]-[07 ] ) and fig.6 , we can now plot the vibrational parameter @xmath117 vs. the equivalent degree of alignment @xmath78(fig.7 ) . due to the second parameter g@xmath7 a family of curves is produced , so that the effective @xmath117 corresponding to a pair ( @xmath136 ) can be found from the plot . knowledge of the efg orientation and sign is not ubiquitous in mssbauer spectroscopy . since the correct attribution of the doublet lines to the @xmath131 and @xmath130 transitions is not always obvious , we plot in fig.7 also the family of curves , which correspond to the false attribution . true attribution regions correspond to the bottom - left ( oblate - platy ) and top - right ( prolate - acicular ) quarters of the plot . false attribution regions correspond to the top - left ( prolate - platy ) and bottom - right ( oblate - acicular ) regions . in other words , the curves with positive slope refer to the true attribution , and the curves with negative slopes refer to the false attribution . in contrast to lnbamn@xmath0fe@xmath1o@xmath2 the left - hand line of the doublet turn to be more intense in the lnbamn@xmath0fe@xmath1o@xmath3 series . the swap between the @xmath131 and @xmath137 lines originates from the reversal of the sign of @xmath138here again the negative sign of @xmath119 is consistent with the ionic point charge model , in which only the charges in first coordination sphere are taken into account@xcite . the case @xmath129 remains unchanged . the origin of asymmetry is therefore attributable to either platy habits of the crystallites , or to oblate vibrational ellipsoids . both of them are expected _ a priory _ and their combined effect must be additive , but not partly extinguishing each other as in above case of lnbamn@xmath0fe@xmath1o@xmath139 . the anisotropies @xmath140 indicated on fig.8 are the `` effective '' values given by the sum of true and equivalent to the pair ( @xmath136 ) values . the negative quadrupole splitting @xmath141 decreases with decreasing the size of ln . this is in agreement with the ionic point charge model@xcite . the absolute value of efg correlates with distortion @xmath142 . in agreement with abrupt change of distortion between sm and nd as shown in fig.3 the absolute value of @xmath141 abruptly drops from 0.7 mm / s to 0.2 mm / s as the ln size increases from sm to nd . the oxygen - saturated `` o@xmath3''-family can be thus divided into two subseries , according to their values of @xmath143 . the spectra of small - size and large size ln - subfamilies are shown in figs . 8 and 9 , respectively . this observation of two subfamilies agrees well with two types of behavior in the phase diagram reported previously@xcite . in the subfamily with small size ln ( y , gd , sm ) the coo is formed by the orbital stacked in sequence aabb along c - axis , and in the the large - size ln - subfamily ( nd , pr ) , the coo is not observed , while at lowering temperature the ferromagnetism first sets in at @xmath144 , followed by the onset at @xmath67@xmath145@xmath144 of antiferromagnetic order of so - called a - type@xcite . mssbauer spectra for the cation - disordered phase ln@xmath100ba@xmath148mn@xmath149fe@xmath147o@xmath24 are shown in fig.10 . iron is in octahedral coordination wherein the small quadrupole splitting can be associated either with the random short - range strain in oxygen sublattice or with the randomness in the ( ba , la)-coordination sphere of the @xmath18fe nuclei . [ ptb ] f10.eps the obtained formulas ( [ 06 ] ) and ( [ 07 ] ) allow to interpret the mssbauer line intensities in terms of the rietveld profile parameters @xmath78 and @xmath79 . in order to compare the rietveld and mssbauer results we present first in fig.11 the contour plot of the doublet asymmetries @xmath150 in function of @xmath78 and @xmath79 . different values of @xmath150 are shown by different colors . total ranges of variation for the alignment degree is @xmath116 and for the fraction of unoriented phase @xmath151 . in fig.11 the ranges @xmath152 and @xmath151 are presented . [ ptb ] f11.eps [ ptb ] f12.eps since all the data in table 1 indicate the platy habits with @xmath153 in all our lnbamn@xmath0fe@xmath1o@xmath2 samples , and with @xmath154 we select these regions in fig.12 , ( a ) and ( b ) , respectively . on the other hand , both these regions correspond to the small enough experimental asymmetries for the `` o@xmath2 '' and `` o@xmath3 '' , respectively . the experimental mssbauer doublet asymmetries from figs . 5 and 8 are marked with the arc segments , and the data from table 1 are indicated by chemical symbol of ln at points with coordinates ( g@xmath80,g@xmath155 ) . interestingly , it turns out that the effects of @xmath78 and @xmath79 on the asymmetry turns out to be approximately equal . the detailed analysis of fig.12 confirms the suggested above picture . in the `` o@xmath2 '' series ( fig.12 , top ) , the asymmetry indicated by symbols is stronger than the asymmetry indicated by arcs . it means that gke reduces the total asymmetry indicated by arcs . in the `` o@xmath3 '' series ( fig.12 , bottom ) , the asymmetry indicated by arcs is stronger than the asymmetry indicated by symbols . it means that gke enhances the total asymmetry indicated by arcs . the effects of texture and gke on spectral asymmetry are only qualitatively considered above to be additive or extinguishing . to express it more exactly , the expression for intensity ratio was formulated@xcite , @xmath156 however , the solutions were yet found either for texture effects , or for gke , separately only . substituting the md function ( eq.1 ) for @xmath87 we propose a general solution for the combined effect of texture and vibrational anisotropy . a replacement of @xmath157 with the minimum texture function @xcite makes the eq . ( 10 ) integrable : @xmath158 here @xmath159 , @xmath160 is the kummer confluent hypergeometric function , @xmath161 and @xmath162 the minimum texture function ( mtf ) compatible with the md function was taken as follows:@xmath163 with @xmath164 the solution ( [ 11 ] ) gives the exact solution only for small asymmetries , because mtf is only a good approximation for the md function for @xmath150 close to 1 . a perfect approximation for the broad range of asymmetries is given elsewhere@xcite . one of the problems , in which the well - oriented powders of platy or acicular crystallites can be useful is the problem of determination of the partial phonon density of states of @xmath18fe in anisotropic materials using the nis spectroscopy@xcite . the vibrational density of states is conventionally derived from measuring the nis spectra in single crystals along different crystal axis . this section presents a proposal for the novel nis experiments using the oriented powder samples . in the nis spectrum of an anisotropic single crystal the phonon dos is weighted by the squared projection of the phonon polarization vectors to the wave vector of the x - ray quantum@xcite . three projected densities of states @xmath165 ( @xmath166 ) of an anisotropic layered material can be determined my measuring the single crystal spectra at three different orientations . in the uniaxially anisotropic material , the powder - averaged nis spectrum is isotropic , however , measuring two spectra on a sample with preferred orientation of crystallites , @xmath167 and @xmath168 at different angles(fig.13 ) provides the full basis for determination of both dos functions @xmath169 and @xmath170 . for this purpose the texture of a sample should be well enough characterized , using rietveld analysis and mssbauer spectroscopy . the simplest experiment that we propose for measuring two components of the uniaxially anisotropic dos on a oriented sample involves the sample stage rotated around the horizontal axis perpendicular to the incident beam ( fig.13 ) . therefore , it is worth to describe the dos in terms of the angles @xmath83 , @xmath171 and the rietveld md function preferred orientation parameters @xmath78 and @xmath79 . [ ptb ] f13.eps let us launch the incident beam under the angle @xmath172 with respect to the preferred axis ( z - axis ) of a platelike or a needle - like crystallite . each dos component is weighted by the projection of the phonon polarization vectors to the wave vector of the x - ray quantum , therefore , the phonon dos projected on the direction defined by the angle @xmath173 can be written : @xmath174 powder averaging of the dos consists in integrating these two terms with the volume of crystallites @xmath175 whose z - axis lies within the cone shell element @xmath176 and result in @xmath177 here the normalization of the texture function to unity and the notation @xmath178 are employed . the polar function @xmath157 , independent on the azimuth angle , is to replace @xmath179 via the coordinate transform from the frame of the beam to the frame of the rotation stage . the ratio of angular elements @xmath180 is @xmath181 @xmath182 and the jacobian of this transform is @xmath181 @xmath183 . using @xmath184 we obtain for the uniaxial symmetry @xmath185 using the inegrated md texture function ( eqs.[06 ] and [ 07 ] ) , we substitute @xmath186 and @xmath187 with @xmath188 and @xmath189 and obtain @xmath190 \cos^{2}\omega+\frac{v(g_{1},g_{2})}{2}\sin^{2}\omega\label{04}%\ ] ] from the eq.([03 ] ) a couple of measurements of dos @xmath191 and @xmath192 at the angles @xmath193 and @xmath194 leads immediately to determination of both dos components : @xmath195 ( \cos^{2}\omega_{1}-\cos^{2}\omega_{2})}\text { \ \ \ } % \ ] ] @xmath196 where @xmath197 and @xmath198 . two aspects of this study were dealt with the methodological and material issues . first , we proposed putting the rietveld analysis in use for nuclear inelastic scattering spectroscopy and developed the method of derivation of the anisotropic phonon dos from the experiments on oriented powder samples . second , in the material aspect of investigating the structure of doped manganites we found a seemingly contradictory events of persistence of the charge - orbital order observed in magnetization and in lattice cell dimension , coexisting with the disordered - state single - site mssbauer spectra in most of the samples , excluding labamn@xmath0fe@xmath1o@xmath2 , and , possibly , smbamn@xmath0fe@xmath1o@xmath2 . however , a plausible interpretation of this combination of results is given suggesting the adaptiveness of the residual short - range coo towards the quenched disorder related to the random distribution of impurities . in oxygen - saturated state ( @xmath12 ) the manganites exhibit the charge and orbital order at ambient temperature for ln = y , gd , sm , but unordered e@xmath199-electronic system for ln = la , pr , nd . fourfold increase of quadrupole splitting was observed in charge and orbitally ordered manganites compared to unordered ones . this is in agreement with the jumplike increase of distortion of reduced perovskite - like cell in the charge and orbitally ordered structures . the light ( i.e. with 2% ) substitution of mn by fe suppresses the temperatures of structural and magnetic transitions by 20 to 50 k. the authors acknowledge the financial support provided through joint jsps - rfbr grant 07 - 02 - 91201 . the work at school of engineering was additionally supported by asahi glass foundation . 99 t. nakajima , h. kageyama , m. ichihara , k. ohoyama , h. yoshizawa , and y. ueda , j. solid st . chem . * 177 * , 987 ( 2004 ) . williams , j.p . attfield , and s.a.t . redfern , phys . rev . * b * * 72 * , 184426 ( 2005 ) . a. i. rykov , europhys . lett . * 85 * ( 2009 ) 16003 . chapman , j.p . attfield , m. molgg , c.m . friend , and t.p . beales , angew . engl . * 35 * , 2482 ( 1996 ) . mcalister , j.p . attfield , j. mater . chem . * 8 * , 1291 ( 1998 ) . f. millange , v. caignaert , b. domengs , and b. raveau , chem . mater . * 10 * , 1974 ( 1998 ) . t. nakajima , h. kageyama , h. yoshizawa , k. ohoyama , and y. ueda , j. phys . . jpn . * 72 * , 3237 ( 2003 ) . t. nakajima , h. yoshizawa and y. ueda , j. phys . jpn . , * 73 * , 2283 ( 2004 ) . d. akahoshi , y. okimoto , m. kubota , r. kumai , t.arima , y. tomioka , and y. tokura , phys . rev . b 70 , 064418 ( 2004 ) . t. nakajima , h. kageyama , y. ueda , j. phys . solids , * 63 * , 913 ( 2002 ) . a. rykov , v. caignaert , and b. raveau , j. solid . chem . * 109 * , 295 ( 1994 ) . kohn , a.i . chumakov , r. rffer , phys . b. * 58 * , 8437 ( 1998 ) . young and d.b . wiles , adv . 24 * , 1 , ( 1981 ) . j. rodriguez - carvajal . physica * b 192 * , 55 ( 1993 ) . f. millange , e. suard , v. caignaert , and b. raveau , mat . bull . * 34 * , 1 ( 1999 ) . v. caignaert , f. millange , b. domengs , and b. raveau , chem . mater . * 11 * , 930 ( 1999 ) . l. er - rakho , c. michel , p. lacorre and b. raveau , j. solid state chem . * 73 * ( 1988 ) 531 . t. nakajima , h. kageyama , h. yoshizawa , and y. ueda , j. phys . jpn * 71 * , 2843 ( 2002 ) d. akahoshi , m. uchida , y. tomioka , t. arima , y. matsui , and y. tokura , phys . lett . * 90 * , 177203 ( 2003 ) . y. ueda and t. nakajima , progr . 35 * , 397 ( 2007 ) . belik , n. hayashi , m. azuma , s. muranaka . , m. takano , and e. takayama - muromachi , j. solid st . chem . * 180 * , 3401 ( 2007 ) . incorporated into fullprof program , http://www-llb.cea.fr/fullweb/powder.htm . dollase , j. appl . cryst . * 19 * ( 1986 ) 267 - 272 . h.m . rietveld , j. appl . cryst . * 2 , * 65 ( 1969 ) . taskin , a.n . lavrov , yoichi ando , progr . chem . * 35 * 481 ( 2007 ) . rykov , y. ueda and k. nomura , arxiv:0902.2027v1 [ cond-mat.str-el ] . shenoy , g.m . friedt , nucl . 136 * , 569 ( 1976 ) . rykov , a. ducouret , n. nguyen , v. caignaert , f. studer and b. raveau , hyperfine interact . * 77 * ( 1993 ) 277 . mombr , k. prassides , c. christides , r.erwin , m. pissas , c. mitros , and d. niarchos , j. phys . : cond . mat . * 10 * , 1247 ( 1998 ) . pfannes and u. gonser , appl . * 1 * , 93 ( 1973 ) . h .- d . pfannes and h. fisher , appl . * 13 * , 317 ( 1977 ) . greneche and f. varret , j. phys . * c15 * , 5333 ( 1982 ) . rykov , m. seto , y. ueda and k. nomura , arxiv:0902.1801v1 [ cond-mat.mtrl-sci ] m. seto , y. yoda , s. kikuta , x.w . zhang and m. ando , phys . lett.*74 * , 3828 ( 1995 ) . fig.1 . the crystal structures and symmetry groups of the layer - ordered lnbamn@xmath7o@xmath2(a ) , lnbamn@xmath7o@xmath3(b ) and disordered ln@xmath148ba@xmath100mno@xmath24 ( c ) . fig.2 . powder x - ray diffraction patterns of the tetragonal phases of gdbamn@xmath0fe@xmath1o@xmath2 and gdbamn@xmath0fe@xmath1o@xmath3 refined using space groups @xmath40 ( no . 129 ) and @xmath41 ( no.123 ) , respectively . lattice parameters of the reduced perovskite - like cell vs. volume of this cell . sm@xmath22nd@xmath23bamn@xmath0fe@xmath1o@xmath17 and sm@xmath101nd@xmath22bamn@xmath0fe@xmath1o@xmath17 are denoted by `` 1 '' and `` 2 '' , respectively . fig.4.magnetic susceptibility @xmath200 measured in the external field @xmath201 of 1 koe per mole of formula units in lnbamn@xmath0fe@xmath1o@xmath3 for ln = y , sm and ( nd@xmath22sm@xmath23 ) . the zero - fied - cooled magnetization was measured at heating the samples up to @xmath32 of 600 k ( ln = y ) , 400 k ( ln = sm ) and 370 k ( ln = nd@xmath22sm@xmath23 ) and then at cooling from @xmath202 . the arrows indicate the temperatures of phase transitions in corresponding pure lnbamn@xmath7o@xmath3 without fe substitution as reported previously@xcite . mossbauer spectra in oriented samples of @xmath18fe - doped oxygen - poor series of the a - site ordered layered manganites lnbamn@xmath0fe@xmath102o@xmath2 . fitted values of isomer shift , quadrupole splitting , and area asymmetry are indicated on the spectra . the spectra are fitted with one or two asymmetric doublets . meltdown of mn@xmath13/mn@xmath10charge order results in single - site spectra for @xmath18fe , while residue of the second doublet is attributed to the remainder of the charge / orbital order . determination of the parameters of texture @xmath78 and vibrational anisotropy @xmath117 starting from the ratio of line intensities of mssbauer doublet . either texture or gke is supposed to be a single source of spectral asymmetry . strength of preferred orientation @xmath78 is plotted vs. spectral asymmetry @xmath203 for four values of the fraction of unoriented phase @xmath79 . the equivalent vibrational anisotropy parameter @xmath117 versus degree of alignment @xmath204at four values of the unoriented phase @xmath79 . the curves with positive slope refer to the true assignment , and the curves with negative slopes refer to the false assignment of the doublet lines to the @xmath130 and @xmath131 transitions . mossbauer spectra in oriented samples of @xmath18fe - doped oxygen - rich series of the charge and orbitally ordered layered manganites lnbamn@xmath21fe@xmath1o@xmath3 . the spectra are fitted with one asymmetric doublet . meltdown of mn@xmath10/mn@xmath14charge order results in single - site spectra for @xmath18fe . fitted values of isomer shift , quadrupole splitting , and area asymmetry are indicated on the spectra . the values of @xmath205 are the `` effective '' anisotropies given by the sum of true and equivalent to the pair ( @xmath136 ) values . mossbauer spectra in the large - size - ln subseries of the oxygen - rich series of the a - site ordered layered manganites lnbamn@xmath0fe@xmath102o@xmath3 showing no coo transition . the spectra are fitted with one symmetric doublet . contour plot of the mssbauer spectra asymmetry @xmath150 versus @xmath207 and @xmath79 ( eqs . [ 06 ] and [ 07 ] ) . the upper half of the plot corresponds to acicular habitus and the lower part of the plot correspond to the platy habitus of the crystals . note that only the region of relatively weak degrees of alignment for acicular habitus is shown , while full range @xmath92 is shown for platy habitus crystallites . the region of relatively weak alignments exhibits symmetric dependence of @xmath150 on @xmath78 and @xmath79 . regions of @xmath150 vs. @xmath78,@xmath79 plot corresponding to our samples of gdbamn@xmath0fe@xmath1o@xmath2 ( top panel ) and gdbamn@xmath0fe@xmath1o@xmath3(bottom panel ) . the arcs show the spectral asymmetries from figs . 5 and 8 and the element symbols indicate the texturing coordinates ( @xmath78,@xmath79 ) of the samples from the table1 .
a - site ordered manganites lnbamn@xmath0fe@xmath1o@xmath2 and lnbamn@xmath0fe@xmath1o@xmath3 are investigated by x - ray full - profile diffraction and mssbauer spectroscopy . powder samples were oriented with preferred orientation of platy crystallites in the plane of sample surface . march - dollase function of preferred orientation was employed in analysing both the rietveld patterns and the mssbauer spectra . combined effects of preffered orientation and vibrational anisotropy on the line area asymmetry of mssbauer doublet are analysed . constructive and destructive interference between the effects of texture and vibrational anisotropy is observed in lnbamn@xmath0fe@xmath1o@xmath3 and lnbamn@xmath0fe@xmath1o@xmath2 , respectively . both series of the manganites show the main axis of electric field gradient perpendicular to layers ( v@xmath4 ) with @xmath5 in oxygen - poor series and @xmath6 in oxygen - rich series . charge - orbital order ( coo ) melting around fe dopants explains the single - site spectra observed for several ln in both `` o@xmath2 '' and `` o@xmath3 '' series , except labamn@xmath0fe@xmath1o@xmath2 . however , the short - range coo persists to be observed in magnetization and in x - ray patterns . specialpagefalse evenheadoddhead evenfootoddfoot
the first measurement of elliptic flow in pb+pb collisions at the large hadron collider ( lhc ) has just been reported @xcite . the elliptic flow coefficient @xmath3 characterizes the momentum anisotropy of final particle emission in non - central heavy - ion collisions relative to the event - plane , defined by the beam direction and the minor axis of the nuclear overlap region in the collision . it describes the efficiency of the medium generated in the collision to generate from an initial spatial deformation of its density distribution an asymmetry in the final momentum distribution , through interactions of the medium constituents . this efficiency increases with the coupling strength between those constituents and becomes maximal for an infinitely strongly coupled medium . in this limit the mean free path of the constituents becomes as small as allowed by the uncertainty relation @xcite , and the medium reaches very quickly a state of approximate local thermal equilibrium which allows to describe its evolution with fluid dynamics . for given initial spatial deformation of the collision fireball , ideal fluid dynamics ( which assumes zero mean free path ) is expected to generate the largest possible elliptic flow @xcite . shear viscosity , a consequence of non - zero mean free paths and limited from below by quantum mechanics @xcite , will lead to a suppression of @xmath3 @xcite . compelling evidence for fluid dynamical behavior of the collision fireballs created in ultrarelativistic heavy - ion collisions , with a very small ratio of shear viscosity to entropy density , @xmath2 , has been found in heavy - ion collisions at the relativistic heavy ion collider ( rhic ) @xcite . the new data from the lhc confirm this picture @xcite and agree , at least qualitatively , with hydrodynamic predictions of elliptic flow for pb+pb collisions at the lhc @xcite . the purpose of the present article is to explore how good this agreement is quantitatively , and to what extent the present and future lhc elliptic flow data can tell us novel facts about the transport behavior of hot qcd matter at temperatures that exceed those accessible at rhic but are within reach at the lhc . similar to the analyses in @xcite , but different from recent hybrid model studies in @xcite , we base our analysis on a purely hydrodynamic approach . while this ignores the fact that the late dilute hadronic stage of the expansion is very dissipative and not well described by fluid dynamics ( neither ideal @xcite nor viscous @xcite ) , the importance of the hadronic phase for the development of elliptic flow is expected to be much reduced at the lhc relative to rhic @xcite . as in refs . @xcite , but different from refs . @xcite , we use _ viscous _ hydrodynamics with a non - zero ( but constant , i.e. temperature independent ) specific shear viscosity @xmath2 , adjusted to spectra and elliptic flow measurements at rhic . our fitted value @xmath4 ( for cgc initial conditions , see below ) is 25% larger than that used by luzum and romatschke @xcite but agrees well with the value for the quark - gluon plasma ( qgp ) viscosity @xmath5 recently extracted from rhic data by using a hybrid viscous hydrodynamic + boltzmann approach ( vishnu @xcite ) . calculations with such a hybrid approach are numerically much more demanding than purely hydrodynamic simulations ; a generalization of the present analysis using vishnu will follow soon and should further improve the reliability of the lhc predictions . in this work , we employ ( 2 + 1)-d viscous hydrodynamics @xcite with the lattice qcd based equation of state s95p - pce @xcite , which accounts for chemical freeze - out before thermal decoupling at @xmath6mev , to simulate the expansion of the collision fireball . from the analysis @xcite of charged hadron spectra and elliptic flow in @xmath7gev au+au collisions at rhic we take over a value of @xmath4 ( corresponding to mc - kln initial conditions , see below ) for the effective specific shear viscosity of the strongly interacting fluid . using the insights obtained from the systematic parameter study presented in @xcite , we initialize the hydrodynamic expansion at time @xmath8fm/@xmath9 and decouple at @xmath10mev at both rhic and lhc energies . for au+au collisions at rhic energies these parameters allow for a good global description of the hadron @xmath11-spectra and differential elliptic flow ( see below ) . lacking strong theoretical or phenomenological guidance how to adjust their values for pb+pb collisions at the lhc , we here decided to keep them unchanged . at thermalization time @xmath12 , we assume that the shear stress tensor is given by its navier - stokes value @xmath13 ( where @xmath14 is the symmetric and traceless velocity shear tensor ) , and that the initial expansion flow is entirely longitudinal with bjorken profile @xmath15 and zero transverse flow velocity . in milne coordinates @xmath16 this corresponds to an initial flow 4-velocity @xmath17 . kinetic freeze - out is implemented by converting the hydrodynamic output to particle spectra with the cooper - frye prescription @xcite on a decoupling surface of constant temperature @xmath18 . we use a quadratic ansatz @xmath19 @xcite for the viscous deviation from local thermal equilibrium of the local phase - space distribution function on the freeze - out surface . our final hadron spectra include decay products from strong decays of all particles and resonances up to 2gev mass @xcite , using the resonance decay program from the azhydro package . for the initial density profile we here use a monte - carlo version @xcite of the kharzeev - levin - nardi @xcite model ( mc - kln ) . the specific implementation used in this work is described in @xcite . the model gives for each event the gluon density distribution in the transverse plane . we assume it to thermalize by time @xmath12 , and convert the gluon density into entropy density @xcite . over one million monte carlo events are re - centered to the beam axis and rotated in the transverse plane such that their minor axis aligns with the impact parameter ( i.e. their `` participant plane '' coincides with the reaction plane ) . after sorting them into collision centrality bins according to their number @xmath20 of participating ( `` wounded '' ) nucleons , we average them to obtain a smooth average initial entropy density which is then converted to energy density using the equation of state . using this smooth energy density as weight , we compute the initial eccentricity @xmath21 and overlap area @xmath22 of the reaction zone ; these represent the corresponding mean values of events in this centrality class . the kln model involves a couple of parameters that need to be adjusted to obtain the correct final charged hadron multiplicity @xmath23 in central au+au collisions at rhic . in @xcite this adjustment was performed for ideal fluid dynamics ( which conserves entropy ) coupled to a hadron cascade . the model then correctly predicts the charged hadron multiplicities at all other collision centralities . in our viscous hydrodynamic model , viscous heating produces additional entropy , leading to somewhat larger final multiplicities . we thus perform an iterative renormalization of the initial entropy density profile until the measured charged hadron multiplicity in the @xmath24 most central @xmath7gev au+au collisions at rhic is once again reproduced . the lower panel of fig . [ f1 ] shows that , after this renormalization , the model again correctly describes the measured @xcite centrality dependence of charged hadron production . the centrality dependence of viscous entropy production ( which is relatively larger in peripheral than in central collisions @xcite ) is ( at least at rhic energies ) sufficiently weak to not destroy the agreement of the model with experimental observations . the ability of the mc - kln model to describe the centrality dependence of charged hadron production without additional parameters is the main reason for choosing it over the mc - glauber model as our basis for extrapolation from rhic to lhc energies . it was recently shown @xcite that this centrality dependence is robust against running coupling corrections @xcite in the balitsky - kovchegov evolution ( on which the kln model is based ) which were found to hardly affect its shape . they do , however , modify the collision energy dependence of particle production , with the lhc pb+pb data being better described if running coupling corrections are included @xcite . our version of the mc - kln model does not include running coupling corrections , and we must normalize the initial entropy density profile for pb+pb collisions at the lhc separately from au+au collisions at rhic . without such an independent renormalization , we overpredict the measured charged multiplicity from central @xmath25tev pb+pb collisions @xcite by about 10% . after renormalization we obtain the solid lines bounding the shaded region in the upper panel of fig . [ f1 ] , with the lower ( upper ) bound corresponding to pb+pb collisions at 2.76 ( 5.5 ) @xmath1gev , respectively . the data in that panel are from the alice collaboration for pb+pb at @xmath25gev @xcite . ( for @xmath26gev pb+pb collisions we assumed @xmath27 ( corresponding to @xmath28 ) , based on an extrapolation of fig . 3 in ref . @xcite . ) one sees that , even without running coupling corrections , but including viscous entropy production , the mc - kln model does a good job in describing the measured centrality dependence of charged hadron production in pb+pb collisions at the lhc . this gives hope that the successful description of the centrality dependence of hadron spectra and elliptic flow at rhic energies ( see below and @xcite ) translates into a reliable prediction of the corresponding centrality dependences in pb+pb collisions at the lhc . figures [ f2 ] and [ f3 ] establish our baseline for the extrapolation to lhc energies . in fig . [ f2 ] we show our purely hydrodynamic fit ( obtained with parameters @xmath12 , @xmath2 , @xmath29 , and @xmath18 set as described abovefm/@xmath9 is 45% smaller than the value of 1.05fm/@xmath9 used for @xmath30 in the vishnu simulations in @xcite . the earlier evolution of hydrodynamic transverse flow arising from this smaller @xmath12 value compensates for the lack of a highly dissipative hadronic phase in the purely hydrodynamic approach . hadronic dissipation leads to a significant broadening in particular of the proton @xmath11-spectra during the hadronic stage which ( given the constraints from the elliptic flow data which prohibit us from simply lowering @xmath18 ) viscous hydrodynamics with temperature - independent @xmath30 can not replicate . ] ) of the hadron spectra measured in @xmath7gev au+au collisions at rhic . [ f2]a shows the mid - rapidity transverse momentum spectra per unit pseudo - rapidity for unidentified charged hadrons from the star @xcite and phenix @xcite experiments compared with the hydrodynamical model . [ f2]b , c show a similar comparison for the @xmath11-spectra per unit rapidity of identified pions and protons from star @xcite and phenix @xcite . in the experimental spectra , protons from weak decays were removed ; star quotes a large systematic error associated with this feeddown correction , and within that large error band the two data sets agree with each other , even if the central values of the star proton data appear to be up to 50% higher than phenix data . our results agree well with the star protons for @xmath31gev/@xmath9 but overpredict the phenix protons by up to a factor 2 . figure [ f3 ] shows the hydrodynamically calculated differential elliptic flow for unidentified charged hadrons in comparison with star @xmath32 data @xcite , for four centrality classes ranging from semi - central to mid - peripheral collisions ( @xmath33 centrality ) . with @xmath34 , viscous hydrodynamics gives an excellent description of the star @xmath35 data , even up to 3gev/@xmath9 in transverse momentum ( i.e. beyond the @xmath11 range where the hydrodynamic description is expected to begin to break down , due to the increasing influence of hard production processes and large uncertainties in the viscous correction @xmath36 to the local phase - space distribution at kinetic freeze - out @xcite ) . looking more carefully one sees that our model slightly overestimates the elliptic flow at low @xmath37gev while underestimating it in the high-@xmath11 region , @xmath38gev . in @xcite we noted a tension in trying to simultaneously fit within a purely viscous hydrodynamic approach the proton @xmath11-spectra and the charged hadron differential elliptic @xmath39 when using eos s95p - pce . even a temperature dependent @xmath40 that has a larger shear viscosity in the hadronic phase could not resolve this tension : in @xcite two of us found that the rhic au+au hadron spectra are insensitive to a temperature - dependent increase of the shear viscosity in the hadron gas phase , as was previously seen in @xcite . [ f2 ] and [ f3 ] demonstrate that this problem is largely resolved when using the @xmath32 data ( fig . [ f3 ] ) instead of @xmath41 ( see fig . [ f8 ] further below ) : we obtain an excellent description of the differential elliptic flow , together with an acceptable description ( within large experimental uncertainties ) of the @xmath11 spectra . overall , the viscous fluid dynamic description of the hadron spectra and charged hadron elliptic flow @xmath42 shown here is of similar quality as the hybrid model description with vishnu presented in @xcite . since purely hydrodynamic simulations are numerically much less costly than calculations with vishnu , we will now use them to generate a broad range of predictions for soft hadron production in pb+pb collisions at the lhc . as discussed above , the extrapolation from rhic to lhc is done keeping @xmath43 and @xmath2 fixed . when comparing the resulting viscous hydrodynamic predictions with experimental data from the recently started lhc heavy - ion collision program , we will search for indications from experiment that would motivate changing these parameters . first results for @xmath11-spectra @xcite as well as both the @xmath11-differential and @xmath11-integrated elliptic flow of unidentified charged hadrons @xcite have already been published and will be compared with the theoretical predictions below . additional experimental information on spectra and elliptic flow of identified hadrons will become available soon ; the relevant hydrodynamic predictions are presented in this section . in fig . [ f4 ] we show the transverse momentum spectra for all charged hadrons , as well as for identified pions and protons , for minimum bias collisions of au+au at rhic and pb+pb at the lhc.(b ) and [ f6 ] and average them . any additional observables , such as the minimum bias elliptic flow in fig . [ f8 ] below , are calculated from these minimum bias spectra . ] for rhic we compare with data from the phenix collaboration @xcite . the upper and lower bounds of the shaded areas are predictions for minimum bias pb+pb collisions at collision energies of 5.5 and 2.76tev per nucleon pair , respectively . the lhc spectra are visibly flatter than at rhic energies , reflecting stronger radial flow . for central collisions ( @xmath24 centrality ) , the fireball lifetime increases from au+au at rhic to pb+pb at lhc by about 19% and 24% , respectively , for 2.76 and @xmath26tev collision energy ; for peripheral collisions at @xmath44 centrality , the corresponding lifetime increases are even larger ( 34% and 41% , respectively ) . the average radial flow velocity increases in central collisions ( @xmath24 centrality ) by 5 and 7% , respectively , and in peripheral collisions ( @xmath44 centrality ) by 9 and 11% . figure [ f5 ] shows the integrated charged hadron elliptic flow @xmath3 as a function of collision centrality for au+au collisions at rhic and pb+pb collisions at the lhc . at rhic energy , our results ( lower red line ) overestimates the star @xmath35 data by about 11% in mid - central collisions , but agrees nicely with @xmath45 except for the most peripheral collisions . from saturating . in addition , viscous effects are stronger in the small fireballs created in peripheral collisions than in the larger central collision fireballs . both effects together cause the theoretical @xmath3 values to decrease sharply at large collision centralities , in apparent conflict with the experimental data . the experimental @xmath41 and @xmath45 measurements are , however , contaminated by non - flow effects , in particular in very peripheral collisions . once non - flow effects are corrected for @xcite , the experimental @xmath3 values decrease at large collision centralities much in the same way as predicted by hydrodynamics . ] at first sight the overprediction of the @xmath11-integrated @xmath35 at rhic is surprising , given the excellent description of the differential elliptic flow @xmath32 shown in fig . the apparent paradox is resolved by observing that the hydrodnamically computed charged hadron @xmath11-spectra shown in fig . [ f2 ] are somewhat harder than measured , thereby giving too much weight in the @xmath11-integral to the range @xmath46gev/@xmath9 where @xmath32 is large . data is fortuitous and should , in fact , not happen since the measured @xmath45 includes a positive contribution from event - by - event @xmath3 fluctuations @xcite while our hydrodynamic calculation yields the average elliptic flow @xmath47 which is smaller . ] at lhc energy ( @xmath48tev ) our integrated @xmath3 lies between @xmath41 and @xmath35 values measured by the alice collaboration @xcite . again , we overpredict the @xmath11-integrated @xmath35 by about @xmath49 . we note that from rhic to lhc the hydrodynamically computed integrated @xmath3 in mid - central to mid - peripheral collisions increases by about 30% , in agreement with the experimental observations . this is due to reduced viscous suppression effects in the larger and denser fireballs created at the lhc and a longer fireball lifetime which allows the momentum flow anisotropy to approach saturation more closely than at lower energies @xcite . in very peripheral collisions , even at lhc energies such a saturation of @xmath3 does not happen ; this is the reason why in fig . [ f5 ] the integrated @xmath3 is seen to decrease at large collision centralities , both at rhic and lhc . in fig . [ f6 ] we present hadron transverse momentum spectra for pb+pb collisions at lhc energies , for a range of collision centralities . in panel ( a ) we compare the hydrodynamic predictions with first data from the alice experiment @xcite . overall , the theoretical description of these experimental data is of similar quality as for au+au collisions at rhic ( see fig . [ f2 ] ) . in the most central collisions , the hydrodynamical model describes the charged hadron spectrum somewhat better than at rhic , whereas in the very peripheral collisions the hydrodynamic spectra are too flat , presumably due to large viscous shear pressure effects . future comparison with the measured spectra at other collision centralities and for identified hadrons , shown here in panels ( b ) and ( c ) as predictions , should shed further light on the origin of the discrepancy in peripheral collisions . figure [ f7]a shows a comparison of the hydrodynamically generated differential @xmath50 for charged hadrons with the alice @xmath35 data @xcite , for four different collision centralities . for the most peripheral of these , we also show the measured @xmath41 for comparison . the hydrodynamic predictions agree nicely with the data at low @xmath51gev/@xmath9 , but overshoot the experimental values by @xmath52 at larger @xmath11 , especially in the more peripheral bins . in the @xmath53 centrality bin , the theoretical prediction happens to agree nicely with @xmath54 even though the latter should be shifted upward by flow fluctuations that are not included in the theoretical calculation . we note that the theoretical overshoot is less severe in the vishnu hybrid model ( see fig . 3 in @xcite ) than in the purely hydrodynamic simulations shown here . this suggests that the excess of @xmath42 over the measured values at @xmath55gev/@xmath9 in fig . [ f7]a may be caused by an inadequate description of the late hadronic stage and its freeze - out . we can summarize figs . [ f2]a , [ f3 ] , [ f5 ] , [ f6 ] , and [ f7]a by observing that the hydrodynamic model overpredicts the @xmath11-integrated charged hadron @xmath3 by @xmath49 at both rhic and lhc , but for different reasons : at rhic the differential elliptic flow @xmath56 is correctly reproduced while the inverse slope of the theoretical @xmath11-spectra is slightly too large , while the lhc @xmath11-spectra are described a bit better ( at least in the most central collisions where published data are available ) but the slope of @xmath56 at the lhc is slightly overpredicted . panels ( b ) and ( c ) of fig . [ f7 ] give predictions for the differential @xmath42 of identified pions and protons . please note the different shape of the proton @xmath42 from that of the pions at low @xmath11 : radial flow pushes the proton elliptic flow to larger values of @xmath11 . comparing the curves for @xmath0 and @xmath26tev , we see that this `` radial push '' of the proton @xmath3 increases with collision energy , so for higher @xmath57 the rise of @xmath42 is shifted to larger transverse momenta , while at fixed @xmath58gev/@xmath9 the proton elliptic flow _ decreases _ with increasing collision energy . this happens only for heavy hadrons but not for the much lighter pions ( see panel ( b ) ) . in figs . [ f8 ] and [ f9 ] we pursue this theme further , by directly comparing the differential elliptic flows at rhic and lhc energies . in fig . [ f8 ] we show results for minimum bias collisions ; the rhic predictions are compared with available data from star @xcite . we see that at low @xmath11 , the elliptic flow for unidentified charged hadrons ( which are strongly pion dominated ) and for identified pions increases from rhic to lhc whereas the opposite is true for protons . at higher @xmath11 ( @xmath59gev/@xmath9 ) , on the other hand , @xmath42 increases for _ both _ pions and protons as we increase the collision energy . [ f9 ] shows this for a few more hadron species , for the @xmath52 and @xmath53 centrality bins : the heavier the hadron , the stronger a push of @xmath3 towards higher @xmath11 is observed . at sufficiently large @xmath11 , @xmath42 is larger at lhc than at rhic for all particle species , but at low @xmath11 this holds only for pions whereas all heavier hadrons show a decrease of @xmath42 from rhic to lhc at fixed @xmath11 . as the hadron rest mass grows , the crossing point where the decrease of @xmath3 at fixed @xmath11 with rising collision energy turns into an increase shifts to larger @xmath11 values . in view of fig . [ f9 ] , the experimental observation @xcite that for unidentified charged hadrons @xmath50 hardly changes at all from rhic to lhc appears accidental : measured at lhc and at rhic is approximately independent of @xmath11 can not be directly used to conclude that @xmath5 does not change from rhic to lhc . if that argument were correct , this ratio should be independent of @xmath11 not only for the sum of all charged hadrons , but also for each identified hadron species separately . our hydrodynamic calculations show that the latter does not hold even if @xmath2 remains unchanged from rhic to lhc . ] the increase of @xmath42 at fixed @xmath11 for pions is balanced by a corresponding decrease for all heavier hadrons leaving , as it happens , no visible net effect once all charged hadrons are lumped together . in refs . @xcite it was argued that a robust method for extracting the qgp shear viscosity is to fit the collision centrality dependence of the eccentricity - scaled charged hadron elliptic flow @xmath60 with a viscous hydrodynamic + hadron cascade hybrid code . in that study it was found that , at fixed collision energy , plotting @xmath60 against the charged hadron multiplicity density per unit overlap area , @xmath61 , yields `` universal '' curves that depend only on the qgp shear viscosity but not on the model for the initial energy density distribution ( in particular its eccentricity ) . in fig . [ f10 ] we show such a plot for @xmath7gev au+au collisions at rhic together with pb+pb collisions at two lhc energies . the four panels show this scaling in terms of distributions in pseudo - rapidity ( @xmath62 , left column ) or rapidity ( @xmath63 , right column ) , and also compare it for our default choice of using the initial _ energy _ density as weight for the calculation of the average eccentricity @xmath64 and overlap area @xmath65 ( top row ) with what one obtains by evaluating these quantities with the initial _ entropy _ density instead ( as is done in refs . @xcite ) ( bottom row ) . we see that , independent of these choices of representation , the universality of @xmath60 vs. @xmath61 or @xmath66 does not carry over to different collision energies ( at least not for the purely hydrodynamic simulations studied in the present work ) : at the same multiplicity density @xmath61 or @xmath66 , more peripheral higher energy collisions produce less elliptic flow per initial eccentricity than more central lower energy collisions . at fixed @xmath30 , the difference between @xmath7gev au+au and @xmath25tev pb+pb collisions ( red circles vs. green upward triangles in fig . [ f10 ] ) is as large as the difference between @xmath67 and @xmath4 for au+au collisions at fixed @xmath68gev ( red circles vs. black squares ) . we note that the tendency in fig . [ f10 ] of higher energy collisions producing less @xmath60 at fixed @xmath61 than lower energy ones contradicts the opposite tendency observed by hirano _ _ in fig . 3 of ref . @xcite where an ideal hydro + hadron cascade hybrid code was employed.gev au+au collisions , our maximal values for @xmath66 in fig . [ f10]b are significantly larger than those shown in fig . 3 of ref . this is due to a lower normalization of the initial entropy density in @xcite , corresponding to @xmath69 instead of our @xmath70 in central au+au collisions ( t. hirano , private communication ) . ] the authors of @xcite presented strong arguments that their observation of larger @xmath60 at fixed @xmath66 in higher energy collisions is not related to their use of a hadron cascade for describing the late hadronic stage . our opposite finding , on the other hand , is supported by the earlier purely hydrodynamic scaling studies presented in the last two works of @xcite whose authors came to the same conclusion as we do here . at present this discrepancy remains unresolved ; we suspect , however , that the origin of the difference between our work and that of hirano _ et al . _ could be in their use of a more realistic ( 3 + 1)-dimensional hydrodynamic evolution @xcite , although in the earlier ideal fluid hydrodynamical studies at the full rhic energy , the differences between boost invariant and non - boost invariant results were small @xcite . possible consequences of the violation of boost - invariance in rhic and lhc heavy - ion collisions are presently being studied @xcite . before moving on , let us comment on the different shape at the high - multiplicity end of the curves shown in fig . [ f10 ] when using entropy instead of energy density as the weight for calculating the initial eccentricity @xmath64 overlap area @xmath65 : it is caused by the different centrality dependence of the energy and entropy density weighted eccentricities in near - central collisions observed in ref . @xcite whose authors showed that in the most central collisions ( where @xmath64 is dominated by event - by - event shape fluctuations ) the entropy - weighted participant eccentricity decreases faster with decreasing impact parameter than the energy - weighted one . shear viscosity is known to suppress the buildup of elliptic flow . naively , the systematic overprediction of @xmath32 in pb+pb collisions at the lhc seen in fig . [ f7]a , together with the excellent description of the same quantity in au+au collisions at rhic seen in fig . [ f3 ] , thus suggests that the fireball liquid formed in lhc collisions might be slightly more viscous ( i.e. possess larger average @xmath2 ) than at rhic energies @xcite . in this section we present some results using a temperature dependent specific shear viscosity , @xmath72 , that were motivated by such considerations . figure [ f11 ] illustrates the following three trial functions explored in this section : @xmath73 here @xmath6mev is the chemical decoupling temperature and stands for the `` transition temperature '' at which the hadronization of quarks and gluons is complete . in principle , the value of @xmath2 should exhibit a minimum near @xmath29 and increase again in the hadronic phase below @xmath29 @xcite . the authors of @xcite pointed out , however , that at the full lhc collision energy of 5.5@xmath1tev the behavior of @xmath2 at temperatures below @xmath29 has very little effect on the final hadron spectra and their elliptic flow . at 2.76@xmath1tev the effect on elliptic flow was moderate , and negligible on the spectra . here we will concentrate on qualitative aspects of effects arising from a temperature dependent growths of @xmath2 in the high temperature region that can be explored at lhc energies but is beyond the reach of rhic , and continue to set @xmath30 at @xmath74 for simplicity . as pointed out in @xcite , the spectra and elliptic flow in au+au collisions at rhic energies are most sensitive to the average value of @xmath2 in the temperature region below 220 - 230mev . we have checked that altering @xmath2 at higher temperatures as shown in fig . [ f11 ] has little influence on the results at rhic energies shown in sec . [ sec2 ] . figure [ f12 ] illustrates the influence of a linear temperature dependence of @xmath2 as in eq . ( [ eq1 ] ) on the centrality dependence of charged hadron production . the solid black line is the same as shown in the upper part of fig . [ f1 ] where it forms the lower bound of the shaded region ; it corresponds to constant @xmath30 and navier - stokes initial conditions for the shear stress tensor , @xmath75 at @xmath76fm/@xmath9 . the dashed and dash - dotted lines in fig . [ f12 ] use @xmath77 with either navier - stokes ( dashed ) or zero ( dash - dotted ) initial conditions for @xmath78 . these last two lines were normalized to the alice point for the @xmath24 most central pb+pb collisions ( @xmath79 @xcite ) , whereas the black line was normalized to our best guess before the alice data became available ( @xmath80 , corresponding to @xmath81 ) . the centrality dependence is then controlled by the predictions from the mc - kln model , modified by viscous entropy production during the hydrodynamic evolution . we see that even a relatively modest temperature dependent increase of @xmath2 in the qgp phase leads to a significantly stronger non - linearity in the dependence of charged particle production on the number of wounded nucleons . the reason is that an increase of @xmath2 with temperature leads to more viscous heating in central collisions ( which probe higher initial temperatures and such larger effective shear viscosities ) than in peripheral ones ( whose initial temperatures are lower ) . since the entropy production rate is given by @xmath82 this effect is stronger for navier - stokes initial conditions ( where @xmath78 is proportional to the velocity shear tensor @xmath83 which at early times diverges like @xmath84 ) than for zero initial shear stress ( where @xmath78 starts from zero and approaches its navier - stokes value @xmath85 only after several relaxation times @xmath86 when , due to its @xmath84 decay , it has already decreased to much smaller values ) . that are generated by viscous heating during the hydrodynamic expansion : constant @xmath30 : @xmath87 ( @xmath24 ) and 33% ( @xmath44 ) ; @xmath77 with @xmath88 : @xmath89 ( @xmath24 ) and 15% ( @xmath44 ) ; @xmath77 with @xmath90 : @xmath91 ( @xmath24 ) and 49% ( @xmath44 ) . ] if one were to postulate the validity of the mc - kln model as the correct description of the initial particle production , the alice data shown in fig . [ f12 ] would exclude a temperature dependence of @xmath2 as given in eqs . ( [ eq1 ] ) and ( [ eq2 ] ) for naver - stokes initial conditions . while we are not prepared to make such a statement on the basis of fig . [ f12 ] alone , we believe that it is important to point out this relatively strong sensitivity of the centrality dependence of @xmath23 to the transport properties of the expanding fireball medium and to emphasize the constraints it thus places on possible models for the qgp shear viscosity . we now turn to the discussion of the influence of a possible temperature dependence of @xmath2 on the charged hadron @xmath11-spectra and elliptic flow . figure [ f13 ] shows lhc predictions for @xmath25tev pb+pb collisions of @xmath92 centrality . to ensure comparability of the different cases studied in this figure we simply normalized the initial entropy density profile such that we always obtain @xmath93 , i.e. the same value that we had obtained before for constant @xmath30 at this centrality . we first note that for constant @xmath30 , we do nt observe any significant difference in the charged hadron spectra and elliptic flow between zero and navier - stokes initialization for @xmath78 . turning to the temperature - dependent parametrizations @xmath94 , we note that for zero initialization of @xmath78 ( solid lines ) our results agree with those reported in @xcite : an increase of @xmath2 at higher qgp temperatures leads to somewhat harder charged hadron @xmath11-spectra ( i.e. somewhat stronger radial flow , caused by the larger tranverse effective pressure gradients at early times ) and a suppression of the differential elliptic flow ( due to an increase of the time - averaged effective shear viscosity of the fluid ) . it is interesting to observe the hierarchy of the curves in fig . [ f13 ] corresponding to the three parametrizations ( [ eq1])([eq3 ] ) : for the @xmath11-spectra , all three @xmath95-dependent viscosities lead to almost identical hardening effects on the spectral slope , while for the differential elliptic flow @xmath50 the curves are ordered not according to the @xmath2-values at the initial central fireball temperature ( see table [ t1 ] ) , but according to their hierarchy in the @xmath96mev range . in fact , the observed magnitudes of the viscous @xmath3 suppression for the three @xmath97 functions suggest that , at this beam energy and collision centrality , the buildup of elliptic flow is dominated by the qgp transport properties at @xmath98mev . ( at rhic energies , the transport properties for @xmath99mev dominate the generation of @xmath3 @xcite . ) cccc @xmath2 model & @xmath100 & @xmath101(@xmath102 ) & @xmath103(mev ) + @xmath30 & .initial central entropy densities @xmath101 and temperatures @xmath103 for the viscous hydrodynamic simulations of @xmath92 centrality pb+pb collisions at the lhc ( @xmath48tev ) shown in fig . the different models for the @xmath95-dependence of @xmath2 are defined in eqs . ( [ eq1])([eq3 ] ) . `` 0 '' stands for @xmath104 at @xmath12 , `` ns '' stands for navier - stokes initialization of the shear stress tensor , @xmath105 at @xmath12 . [ t1 ] [ cols="^ " , ] + for navier - stokes initial conditions ( dashed lines in fig . [ f13 ] ) , the increase in radial flow caused by an increase of @xmath2 at high temperature is stronger and the viscous @xmath3 suppression is weaker than for zero initial @xmath78 . this is caused by the much larger initial shear stress tensor components in the ns case , compared to the case of @xmath104 where @xmath78 approaches its ( by that time already much smaller ) navier - stokes limit only after several relaxation times @xmath86 @xcite . the increase of @xmath2 with temperature generates a steeper initial transverse effective pressure gradient ( since @xmath78 grows faster than the entropy density @xmath106 when @xmath2 increases with temperature ) , and this generates stronger radial flow . it also causes a larger spatial eccentricity of the initial effective pressure profile which ( when compared to the case of @xmath104 ) generates stronger elliptic flow . in fact , we found that for earlier starting times @xmath12 ( where the navier - stokes values for @xmath78 are even larger ) , the quadratic parametrization @xmath107 with ns initial conditions can lead to _ more _ elliptic flow than a constant @xmath30 , in spite of the larger mean viscosity of the fluid . we conclude from this exercise that a firm determination whether or not the alice data point towards a temperature - dependent growth of @xmath2 with increasing @xmath95 , as expected from perturbative qcd @xcite and ( perhaps ) from lattice qcd @xcite , is not possible without a better understanding of the initial conditions for the energy momentum tensor ( in particular the shear stress components ) at the beginning of the hydrodynamic evolution . whereas generically larger viscosities cause a suppression of the elliptic flow , temperature - dependent viscosities can influence the initial effective pressure profile and its eccentricity in a way that counteracts this tendency and , for some models such as navier - stokes initial conditions , can even overcompensate it . based on an successful global fit of soft hadron production data in @xmath7gev au+au collisions at rhic with a pure viscous hydrodynamic model with cooper - frye freeze - out , presented in sec . [ sec2 ] , we generated hydrodynamic predictions for the @xmath11-spectra and differential elliptic flow of unidentified charged hadrons and identified pions and protons for pb+pb collisions at the lhc . where available , these predictions were compared with available experimental data from the alice collaboration . our extrapolation from rhic to lhc energies was based on the assumption that the qgp shear viscosity @xmath2 does not change with increasing fireball temperature and stays fixed at the value @xmath30 extracted from the rhic data , assuming mc - kln initial conditions . the start time @xmath12 for the hydrodynamic evolution and the freeze - out temperature @xmath18 were held fixed , too . we found that , using the beam energy scaling implicit in the mc - kln model , such an extrapolation gives a good description of the centrality dependence of charged hadron production and the charged hadron @xmath11-spectra in central pb - pb collisions , but overpredicts the slope of the @xmath11-differential elliptic flow and the value of its @xmath11-integrated value by about @xmath49 in mid - central to mid - peripheral collisions . in the most peripheral collisions , the predicted charged hadron @xmath11-spectra are too flat , and the integrated elliptic flow is too small compared to the experimental data . a preliminary study of possible temperature dependent variations of @xmath2 in the high - temperature region explored for the first time at the lhc remained unconclusive but pointed to a clear need for better theoretical control over the initial conditions for the hydrodynamic energy - momentum tensor , in particular its shear stress components . the development of detailed dynamical models for the pre - thermal evolution of the collision fireball and their matching to the viscous hydrodynamic stage is a matter of priority for continued progress towards quantifying the transport properties of the quark - gluon plasma at different temperatures and densities . u. heinz and p. f. kolb , nucl . a * 702 * , 269 ( 2002 ) . g. policastro , d. t. son and a. o. starinets , phys . lett . * 87 * , 081601 ( 2001 ) ; p. k. kovtun , d. t. son and a. o. starinets , phys . rev . lett . * 94 * , 111601 ( 2005 ) . u. heinz and s. m. h. wong , phys . rev . * c66 * , 014907 ( 2002 ) . d. teaney , phys . c * 68 * , 034913 ( 2003 ) . i. arsene _ et al . _ ( brahms collaboration ) , nucl . * a757 * , 1 ( 2005 ) ; b. b. back _ et al . _ ( phobos collaboration ) , _ ibid . , _ p.28 ; j. adams _ et al . _ ( star collaboration ) , _ ibid . , _ p.102 ; k. adcox _ et al . _ ( phenix collaboration ) , _ ibid . , _ r. a. lacey and a. taranenko , pos * cfrnc2006 * , 021 ( 2006 ) ; r. a. lacey _ et al . _ , phys . lett . * 98 * , 092301 ( 2007 ) ; a. adare _ et al . _ , phys . lett . * 98 * , 172301 ( 2007 ) ; h .- j . drescher , a. dumitru , c. gombeaud , and j .- y . ollitrault , phys . c * 76 * , 024905 ( 2007 ) . k. dusling and d. teaney , phys . c * 77 * , 034905 ( 2008 ) ; z. xu , c. greiner , and h. stcker , phys . * 101 * , 082302 ( 2008 ) ; d. molnar and p. huovinen , j. phys . g * 35 * , 104125 ( 2008 ) ; r. a. lacey , a. taranenko and r. wei , in _ proc . 25th winter workshop on nuclear dynamics _ , w. bauer , r. bellwied , and j.w . harris ( eds . ) , ( ep systema , budapest , 2009 ) p. 73 [ arxiv:0905.4368 [ nucl - ex ] ] ; k. dusling , g. d. moore , and d. teaney , phys . c * 81 * , 034907 ( 2010 ) ; a. k. chaudhuri , j. phys . g * 37 * , 075011 ( 2010 ) ; r. a. lacey _ et al . _ , phys . c * 82 * , 034910 ( 2010 ) . p. romatschke and u. romatschke , phys . * 99 * , 172301 ( 2007 ) ; m. luzum and p. romatschke , phys . c * 78 * , 034915 ( 2008 ) . h. song , s. a. bass , u. heinz , t. hirano and c. shen , phys . * 106 * , 192301 ( 2011 ) ; and phys . c * 83 * , 054910 ( 2011 ) . m. luzum , phys . c * 83 * , 044911 ( 2011 ) . r. a. lacey , a. taranenko , n. n. ajitanand and j. m. alexander , phys . c * 83 * , 031901 ( 2011 ) . n. armesto _ et al . _ , j. phys . g * 35 * , 054001 ( 2008 ) . h. niemi , k. j. eskola and p. v. ruuskanen , phys . c * 79 * , 024903 ( 2009 ) g. kestin and u. heinz , eur . j. c * 61 * , 545 ( 2009 ) ; u. heinz , nucl . a830 * , 287c ( 2009 ) . m. luzum and p. romatschke , phys . rev . lett . * 103 * , 262302 ( 2009 ) . p. bozek , m. chojnacki , w. florkowski and b. tomasik , phys . lett . * b694 * , 238 ( 2010 ) ; p. bozek , _ ibid . _ * b699 * , 283 ( 2011 ) . t. hirano , p. huovinen and y. nara , phys . c * 83 * , 021902 ( 2011 ) . b. schenke , s. jeon and c. gale , phys . lett . * b702 * , 59 ( 2011 ) . h. song , s. a. bass and u. heinz , phys . c * 83 * , 054912 ( 2011 ) . t. hirano , u. heinz , d. kharzeev , r. lacey and y. nara , phys . b636 * , 299 ( 2006 ) . h. song , s. a. bass and u. heinz , phys . c * 83 * , 024912 ( 2011 ) . t. hirano , u. heinz , d. kharzeev , r. lacey and y. nara , j. phys . g * 34 * , s879 ( 2007 ) . h. song and u. heinz , phys . * b658 * , 279 ( 2008 ) ; phys . c * 77 * , 064901 ( 2008 ) ; * 78 * , 024902 ( 2008 ) ; h. song , ph.d thesis , the ohio state university ( august 2009 ) [ arxiv:0908.3656 [ nucl - th ] ] . p. huovinen and p. petreczky , nucl . * a837 * , 26 ( 2010 ) . c. shen , u. heinz , p. huovinen and h. song , phys . c * 82 * , 054904 ( 2010 ) . f. cooper and g. frye , phys . d * 10 * , 186 ( 1974 ) . j. sollfrank , p. koch and u. heinz , phys . * b252 * , 256 ( 1990 ) ; z. phys . c * 52 * , 593 ( 1991 ) ; j. sollfrank , p. huovinen , m. kataja , p. v. ruuskanen , m. prakash and r. venugopalan , phys . c * 55 * , 392 ( 1997 ) . a. adil , h. j. drescher , a. dumitru , a. hayashigaki and y. nara , phys . c * 74 * , 044905 ( 2006 ) . h. j. drescher and y. nara , phys . c * 76 * , 041903 ( 2007 ) . d. kharzeev , m. nardi , phys . * b507 * , 121 ( 2001 ) ; d. kharzeev , e. levin , phys . * b523 * , 79 ( 2001 ) . t. hirano and y. nara , phys . c * 79 * , 064904 ( 2009 ) ; and nucl . phys . * a830 * , 191c ( 2009 ) . z. qiu and u. heinz , phys . c * 84 * , 024911 ( 2011 ) . b. b. back _ et al . _ [ phobos collaboration ] , phys . c * 70 * , 021902 ( 2004 ) ; b. alver _ et al . _ [ phobos collaboration ] , _ ibid . _ * 80 * , 011901 ( 2009 ) . j. l. albacete and a. dumitru , arxiv:1011.5161 [ hep - ph ] . i. balitsky , phys . rev . d * 75 * , 014001 ( 2007 ) . j. l. albacete , phys . rev . * 99 * , 262301 ( 2007 ) . y. v. kovchegov and h. weigert , nucl . * a784 * , 188 ( 2007 ) ; nucl . phys . * a807 * , 158 ( 2008 ) . k. aamodt _ et al . _ [ alice collaboration ] , phys . lett . * 106 * , 032301 ( 2011 ) . k. aamodt _ et al . _ [ alice collaboration ] , phys . lett . * 105 * , 252301 ( 2010 ) . j. adams _ et al . _ [ star collaboration ] , phys . rev . lett . * 91 * , 172302 ( 2003 ) . s. s. adler _ et al . _ [ phenix collaboration ] , phys . c * 69 * , 034910 ( 2004 ) . j. adams _ et al . _ [ star collaboration ] , phys . lett . * 92 * , 112301 ( 2004 ) ; b. i. abelev _ et al . _ [ star collaboration ] , _ ibid . _ * 97 * , 152301 ( 2006 ) . b. i. abelev _ et al . _ [ star collaboration ] , phys . c * 79 * , 034909 ( 2009 ) . s. s. adler _ et al . _ [ phenix collaboration ] , phys . c * 69 * , 034909 ( 2004 ) . y. bai , ph.d . thesis , nikhef and utrecht university , the netherlands ( 2007 ) . b. i. abelev _ et al . _ [ star collaboration ] , phys . c * 77 * , 054901 ( 2008 ) . c. shen and u. heinz , phys . c * 83 * , 044909 ( 2011 ) . h. niemi , g. s. denicol , p. huovinen , e. molnar and d. h. rischke , phys . lett . * 106 * , 212302 ( 2011 ) . k. aamodt _ et al . _ [ alice collaboration ] , phys . * b696 * , 30 ( 2011 ) . k. aamodt _ et al . _ [ alice collaboration ] , phys . lett . * 107 * , 032301 ( 2011 ) ; r. snellings _ et al . _ [ alice collaboration ] , arxiv:1106.6284 [ nucl - ex ] . j. y. ollitrault , a. m. poskanzer and s. a. voloshin , phys . c * 80 * , 014904 ( 2009 ) . p. f. kolb and u. heinz , in _ quark - gluon plasma 3 _ , edited by r. c. hwa and x .- n . wang ( world scientific , singapore , 2004 ) , p. 634 [ arxiv : nucl - th/0305084 ] ; j. adams _ et al . _ [ star collaboration ] , phys . c * 72 * , 014904 ( 2005 ) . c. shen , b. schenke , and u. heinz , in preparation . t. hirano and k. tsuda , phys . c * 66 * , 054905 ( 2002 ) . l. p. csernai , j. i. kapusta and l. d. mclerran , phys . * 97 * , 152303 ( 2006 ) j. w. chen , m. huang , y. h. li , e. nakano and d. l. yang , phys . * b670 * , 18 ( 2008 ) . j. i. kapusta , in _ relativistic heavy ion physics _ , landolt - brnstein new series , vol . i/23 , edited by r. stock , ( springer verlag , new york , 2010 ) [ arxiv:0809.3746 [ nucl - th ] ] . b. arnold , g. d. moore and l. g. yaffe , jhep * 0305 * , 051 ( 2003 ) . h. b. meyer , phys . d * 76 * , 101701 ( 2007 ) .
a comprehensive viscous hydrodynamic fit of spectra and elliptic flow for charged hadrons and identified pions and protons from au+au collisions of all centralities measured at the relativistic heavy ion collider is performed and used as the basis for predicting the analogous observables for pb+pb collisions at the large hadron collider at @xmath0 and 5.5@xmath1tev . comparison with recent measurements of the elliptic flow of charged hadrons by the alice experiment shows that the model slightly over - predicts the data if the same ( constant ) specific shear viscosity @xmath2 is assumed at both collision energies . in spite of differences in our assumptions for the equation of state , the freeze - out temperature , the chemical composition at freeze - out , and the starting time for the hydrodynamic evolution , our results agree remarkably well with those of luzum [ m. luzum , phys . rev . c * 83 * , 044911 ( 2011 ) ] , indicating robustness of the hydrodynamic model extrapolations . future measurements of the centrality and transverse momentum dependence of spectra and elliptic flow for identified hadrons predicted here will further test the model and shed light on possible variations of the quark - gluon transport coefficients between rhic and lhc energies .
solar flares are sudden releases of magnetic energy which occur in the solar atmosphere , reaching up to 10@xmath0 ergs and lasting anything from a few to a few tens of minutes ( see , e.g. , @xcite for a recent comprehensive review ) . flare emission is detected in all em bands , from radio to gamma - rays , and also are detected in the variations of the flux density of solar energetic particles approaching the earth with energies ranging from a few tens of kev to several gev . similar events are observed on solar type stars , as well as on stars of other classes , e.g. of the uv cet type . solar flares , being among the most powerful and most energetic physical phenomena in the solar system , attract attention because of their decisive role in solar - terrestrial relations and space weather , their connection with the earth s climate , as well as in the context of basic plasma physics research . often , the em radiation generated in solar and stellar flares shows a pronounced oscillatory pattern , with characteristic periods ranging from a fraction of a second to several minutes . the oscillations can be seen in all observational bands , often in phase , and the modulation depth sometimes exceeds 100% ( see , e.g. @xcite ) . traditionally , these oscillations are referred to as quasi - periodic pulsations ( qpp ) , to emphasise that they often contain apparent amplitude and period modulation , and that their shape is often anharmonic . interest in solar flare qpp is connected first of all with the possible role these oscillations play in flare energy releases . indeed , there must be some physical reason for the flaring emission being arranged in a sequence of periodic bursts in a significant fraction of observed flares . the occurrence of qpp put additional constraints on the interpretation and understanding of the basic processes operating in flares , such as particle acceleration , magnetic energy liberation and plasma hydrodynamics and thermodynamics . also , observed parameters of the oscillation ( e.g. the periods , modulation , typical signatures , spatial information ) should be connected with the physical parameters of the flaring plasma and hence can be used for diagnostic purposes through the method of coronal seismology ( see @xcite for more detail ) . the qpp - based seismology has a number of attractive advantages . the high level of the emitted power allows observers to increase the time resolution of their instruments , easily resolving e.g. the alfvn wave transverse transit time and perhaps even the ion gyroperiod . the flare excites various kinds of magnetohydrodynamic ( mhd ) waves and oscillations in the compressible and elastic surrounding plasmas , which can be detected in the modulation of the flaring emission . moreover , qpp are also observed in stellar flares , and this opens up interesting perspectives for stellar coronal seismology and various comparative studies . kev , thin solid curve ) and the song 80 - 200 kev count rate ( thick solid curve ) . panel ( b ) : the rhessi corrected count rate in the channels 25 - 50 kev ( solid ) , 50 - 100 kev ( dotted ) , and 100 - 300 kev ( dashed ) . ] mathematically , in a number of cases , flaring qpp can be considered as an auto - oscillation : a undamped oscillation in a non - linear dynamical system , whose amplitude and frequency are largely independent of the initial conditions , and are determined by the properties of the system itself . dynamical systems capable of performing auto - oscillations include clocks , generators of electric vibrations , wind and string musical instruments , etc . many laboratory and natural plasma systems are observed to exhibit self - sustaining oscillatory phenomena , which can be classified as auto - oscillations . plasma physics examples of auto - oscillations are the sawtooth oscillations , periodic shedding of hydrodynamic or alfvnic vortices in the interaction of magnetohydrodynamic flows with stationary obstacles , oscillations of the magnetic reconnection rate and the current channel in laboratory experiments ( e.g. @xcite ) , and nonlinear thermal over - stability of the plasma @xcite . investigation of plasma auto - oscillations is a rapidly developing research topic , and the study of qpp in solar flares provide us with unique and abundant information of fundamental importance . here we present a brief review of the current trends in the observational and theoretical study of qpp in solar flares , concentrating on their auto - oscillatory nature . comprehensive discussion of this topic , including alternative theories , such as the modification of the non - thermal particle dynamics by a sausage mode and oscillatory regimes of wave - particle interaction , can be found in @xcite . often , qpp can be clearly seen in the time series ( light curves " ) obtained in the observational channels , associated with non - thermal electrons , accelerated in the flare . figure [ fig : l ] shows a typical example of light curves obtained in the microwave and hard x - ray bands with four different instruments during a powerful solar flare ( of the goes - class x1.2 ) . the independent detection of qpp with different ground - based and space - borne instruments eliminates the possibility that they are non - solar in origin . a detailed phenomenological study of this qpp event was recently presented in @xcite . at least four one - minute qpp of growing amplitude are visible between 01:01 and 01:05 ut . a number of similar events are summarised in the on - line archive which provides a list of currently known solar qpp events . an interesting feature of the qpp is its apparent anharmonic shape ( see section [ sec : emd ] ) . recently , it was discovered that in some qpp several different oscillatory patterns occur simultaneously . in weak flares , qpp can be seen in , e.g. , soft x - ray emission @xcite . however , in strong flares , qpp are observed at photon energies up to 2 - 6 mev , which are associated with accelerated ions @xcite , as well as at the lower energies associated with non - thermal electrons . this indicates that the likely cause of the qpp is the time variability of the charged particle acceleration process , e.g. a periodic regime of magnetic reconnection . the mechanism responsible for the fast release of magnetic energy in the solar corona , and its conversion into heat , and kinetic energy of bulk flows and accelerated charged particles is believed to be magnetic reconnection . details of this mechanism are still under intensive investigation ; although it is clear that the process is essentially non - stationary : the steady supply of the magnetic energy by reconnection inflows results in bursty energy release . there are two possibilities for this process to be periodic : spontaneous , when the periodicity is determined by the plasma parameters in the reconnection site , and periodically triggered , when the periodicity is prescribed externally , e.g. by an external resonator . spontaneous periodic or quasi - periodic release of energy by magnetic reconnection can be considered as a load / unload or relaxation process . we may illustrate it by a dripping " model @xcite : slowly and continuously leaking water is gathering on the ceiling , forming a bulge of growing mass , and when the gravitational force becomes sufficiently strong to counteract the surface tension force , the droplet reconnects from the rest of the water and falls down . then the situation repeats , and the dripping rate can be quite periodic and stable . the period of dripping is determined by the inflow rate , and the gravitational and surface tension forces . similarly , in the case of magnetic reconnection , the magnetic energy can be continuously supplied by inflow , build up in the vicinity of a flare epicentre , and , when a certain critical level of the magnetic field complexity is reached , the energy is released by a burst , and the buildup of the energy repeats again . theoretically , a periodic regime of magnetic reconnection can be achieved in several physical situations . one possibility is the coalescence of two magnetic flux tubes , e.g. , a collision of two twisted coronal loops @xcite . the induced oscillation is essentially anharmonic : the magnetic field strength experiences periodic sudden increases , causing periodic spikes of the current density @xcite . the period is estimated as a product of the plasma-@xmath1 and the transverse alfvn transit time across the region of the interaction ( which is definitely smaller than the minor radii of the colliding loops ) . more generally , the period is determined by the plasma @xmath1 , the magnetic twist ( the ratio between the poloidal and toroidal components of the field in the loops ) , and the colliding velocity of the loops . according to numerical simulations , the oscillations can have double or triple sub - peaks , which is consistent with some observational examples of qpp @xcite . during magnetic reconnection , steep gradients of the magnetic field produce high values of the electric current density , which can reach the threshold of plasma micro - instabilities and hence cause the switching from classical to anomalous resistivity . the positive feedback between reconnection , plasma acceleration and rise of the resistivity can result in a periodic regime of magnetic reconnection @xcite . the periodicity is manifested through the repeated generation of magnetic islands or plasmoids , and their coalescence . the principle effect that leads to the repetitive reconnection is the inability of the plasma to carry a sufficient amount of magnetic flux into the diffusion region to support the alfvnic outflow in the steady petchek state . thus the system keeps switching between petchek and sweet parker regimes . the period of the repetition was found to be about thirteen alfvn crossing times between neighbouring magnetic x - points in the generated chain of plasmoids . generation of similar sequences ( or chains ) of plasmoids in a reconnecting current sheet has also been observed in the 2d numerical resistive - mhd experiments with large lundquist numbers @xcite . this effect can explain the oscillatory reconnection in the numerical experiments on the emergence of a magnetic flux rope into the solar atmosphere endowed with a vertical magnetic field @xcite . a series of reconnection reversals with the period of several minutes was observed , whereby reconnection occurred in distinct bursts and the inflow and outflow magnetic fields of one burst became the outflow and inflow fields in the following burst , respectively . during each burst the gas pressure in the bounded outflow regions increases above the level of that in the inflow regions and , eventually , gives rise to a reconnection reversal . nonlinear coupling of the tearing mode and the kelvin helmholtz instabilities was shown to have a periodic ( or over - stable ) regime @xcite too . the most preferable conditions for this regime appear in high @xmath1 plasmas . in 2.5d compressible visco - resistive mhd simulations , the period of the induced oscillations was found to be about 50 alfvn transit times across the current sheet half - width . the vicinity of a magnetic x - point , one of the preferable sites for magnetic reconnection , is a fast magnetoacoustic resonator . standing fast waves trapped in the resonator modify the magnetic field and plasma density , and hence affect periodically the reconnection rate . the oscillatory regime of reconnection , caused by an @xmath2 oscillation in a 2d x - type neutral magnetic point , has been found in the numerical resistive - mhd simulations . the nonlinear compressible fast - mode oscillation deforms the x - point , causing periodic creation of reconnecting current sheets . the plane of the sheets alternates between vertical and horizontal orientations . the principle reason for the oscillation is the inertial overshooting of the plasma , which carries more flux through the neutral point than is required for static equilibrium @xcite . in terms of the dripping " model discussed in section [ sec : spon ] , the periodically triggered regime of reconnection can be seen as periodic shaking of the ceiling . in this case , there is a certain region on the parametric plane showing the period and the amplitude of the shaking force , which corresponds to the dripping rate coinciding with the period of the shaking . in mhd , the external shaking can be associated with an mhd oscillation or wave generated outside the flare epicentre . the link of the mhd wave and the reconnection rate can be achieved by several mechanisms . one option is if there is an oscillating plasma structure near the potential reconnection site . transverse oscillations create periodic fast magnetoacoustic wave , approaching the reconnection site . the interaction of the fast wave with , e.g. , a magnetic x - point , is accompanied with a periodic creation of very localised and sharp spikes of the electric current density . the spike acts as a seed for the onset of current - driven plasma micro - instabilities , causing a sudden increase in the plasma resistivity in the vicinity of the x - point and hence periodically triggering magnetic reconnection . another mechanism , proposed in @xcite , links the periodic triggering of reconnection by a slow magnetoacoustic wave : the wave periodically perturb the plasma density in the reconnection site , modulating its rate . such a relationship between three - minute oscillations in sunspot atmospheres , interpreted as slow waves , and qpp in solar flares in an adjacent coronal active region was observationally established in . the complexity of the physics associated with the generation of flaring qpp by oscillatory regimes of magnetic reconnection , discussed above , makes the flaring light curves intrinsically non - stationary and nonlinear . hence , proper analysis of light curves exhibiting qpp should take into account that feature . the traditional approach to the detection of qpp is based upon various versions of fourier analysis , or on wavelet analysis with the morlet mother function ( e.g. ) , which both consider the signal as a sum of linear harmonic functions . harmonic functions , being the eigenfunctions of linear oscillatory systems , are definitely not the most suitable basis for the decomposition of nonlinear and non - stationary series and revealing the physical mechanisms responsible for their time - variability . a recently designed method of empirical mode decomposition ( emd , @xcite ) is a promising alternative to the fourier - based techniques in the study of flaring qpp . emd decomposes a signal into a small number of intrinsic mode functions . an intrinsic mode function satisfies the following conditions : in the whole data set , the number of extrema and the number of zero crossings must either equal or differ at most by one , and at any data point , the mean value of the envelope defined using the local maxima and the envelope defined using the local minima is zero . thus , emd does not require the intrinsic mode functions to be harmonic , allowing for the extraction of signals of an anharmonic shape ( e.g. such as saw - tooth , square , triangular ) . in application to solar coronal oscillations , the first use of emd was in @xcite . the first application of emd to solar flare light curves was in @xcite . figure [ fig:2 ] shows the emd - filtering of the light curves of the flare shown in figure [ fig : l ] . emd determined eight intrinsic modes . after removal of the apparent high - frequency noise and the long - durational trend , the synthesised signal keeps the anharmonicity of the oscillations , pointed out in section [ sec : obs ] . similarly , emd can be used as an adaptive filtering that naturally preserves the period and amplitude modulation in the signal . observational properties of solar flare qpp , the modulation depth , the anharmonic shape and the period , suggest their auto - oscillatory nature and the possible relationship with oscillatory regimes of magnetic reconnection . those regimes have been observed in several numerical mhd experiments , and can occur spontaneously or be periodically triggered by an externally generated mhd oscillation or wave . the study of solar and stellar flare qpp opens up very interesting perspectives for revealing the basic physical processes operating in flares and the mechanisms for magnetic reconnection , charged - particle acceleration and energy deposition . an interesting avenue in the theoretical investigation would be creation of a low - dimensional model linking observable parameters of qpp with the physical conditions in the reconnection site . in particular , it is necessary to understand whether the oscillatory regime is associated with relaxation oscillations , or is better described by some kind of a limit - cycle dynamics . such a model would become a basis for plasma diagnostics with the use of qpp . in data analysis , a prospective novel approach is the application of the emd technique that allows one to extract nonlinear and non - stationary properties of qpp . the primary questions to be answered observationally are whether qpp is an intrinsic feature of flaring energy releases , what the typical shape of the oscillatory pattern is , the phenomenological link between the main periods and other observables , the existence of multi - periodic regimes and the typical period ratios . all this makes the study of solar flare qpp an interesting , promising and rapidly developing novel research area , which will shed light not only on the physical processes operating in the solar and stellar atmospheres , but also of fundamental importance for basic plasma physics . the work was supported by the royal society uk - russian international joint project , stfc , and the grants rfbr 10 - 02 - 01285-a and 10 - 02 - 00153-a .
electromagnetic ( radio , visible - light , uv , euv , x - ray and gamma - ray ) emission generated by solar and stellar flares often contains pronounced quasi - periodic pulsations ( qpp ) . physical mechanisms responsible for the generation of long - period qpp ( with the periods longer than one second ) are likely to be associated with mhd processes . the observed modulation depths , periods and anharmonicity of qpp suggest that they can be linked with some kind of mhd auto - oscillations , e.g. an oscillatory regime of magnetic reconnection . such regimes , of both spontaneous and induced nature , have been observed in resistive - mhd numerical simulations . the oscillations are essentially nonlinear and non - stationary . we demonstrate that a promising novel method for their analysis is the empirical mode decomposition technique .
the integer quantum hall effect ( iqhe ) is observed in two - dimensional ( 2d ) electron systems , in the presence of a strong perpendicular magnetic field.@xcite current understanding of iqhe is mostly based on studies of non - interacting electron models , with the simplifying assumption that electron - electron interaction does not play an important role . indeed , the phenomenology of iqhe may be explained very well by the non - interacting electron model in a strong magnetic field : the magnetic field gives rise to landau quantization and the large gap necessary for dissipationless transport , while disorder localizes electronic states away from the centers of landau bands , such that the hall conductance remains quantized as the fermi energy lies in the localized region of the density of states ( dos ) ; transitions between different quantum hall phases occur when the fermi energy sweeps through critical energies in the middle of landau bands , where states are delocalized ( i.e. , have a diverging localization length).@xcite the scaling theory of localization@xcite predicts that all states in a system of noninteracting electrons are localized in 2d , in the absence of a magnetic field . therefore a 2d electron gas system must undergo a phase transition ( or a series of phase transitions ) from a quantum hall state to the insulating phase , as the strength of the magnetic field ( @xmath0 ) goes to zero.@xcite the nature of such transition(s ) is of current interest,@xcite and is the subject of the present paper . on the theoretical front , khmelnitskii@xcite and laughlin@xcite proposed , based on semiclassical considerations , that delocalized states at centers of landau bands float " upward as @xmath0 decreases and mixing between different landau bands becomes important , and disappear at infinite energy in the @xmath1 limit , thus reconciling the very different behavior at zero and strong magnetic fields . this simple and elegant picture is the basis of the global phase diagram of quantum hall effect proposed by kivelson , lee and zhang,@xcite and has received support from both experimental@xcite and microscopic theoretical@xcite studies . on the other hand , based on numerical studies of a one - band tight binding model on a square lattice , authors of refs . ( see also refs . ) concluded that instead of floating to infinite energy in the @xmath1 limit , the extended states disappear at finite @xmath0 , and proposed a very different phase diagram . on the experimental side , the situation is also rather unclear . although most experiments are consistent with the global phase diagram , people have routinely observed direct transition from quantum hall phases with @xmath2 to an insulator as @xmath0 decreases , where @xmath3 is the hall conductance in unit of @xmath4 . this is in apparent contradiction with the global phase diagram , which predicts that @xmath3 can only change by @xmath5 at each integer quantum hall transition . some of these discrepancies , like the @xmath6 transitions,@xcite may be explained as due to the small zeeman splitting at low @xmath0 , which gives rise to spin - unresolved quantum hall transitions . the recently observed @xmath7 transition,@xcite however , is apparently quite difficult to interpret along the same line . most of the previous theoretical studies focus on the continuum model . it was recently argued,@xcite however , that the existence of an underlying lattice ( and the corresponding conduction band ) is important to understand the transition between a quantum hall state and the insulating state at low @xmath0 , even though the band width is much bigger than the fermi energy and landau level spacings . a crucial difference between the continuum model and tight - binding models is that in the latter case , there exist current carrying states that carry negative hall conductance near the band center ( and the total hall conductance is zero for the entire band ) , while such states do not exist in the former case . however it remains unclear how to interpolate between the lattice and continuum models . in this paper we present new results of numerical studies of the tight binding lattice model in the presence of a uniform magnetic field and random potential . unlike previous studies@xcite which have been focusing on special cases where there is particle - hole symmetry and critical energies carrying negative hall conductance only exist in the central band(s ) , we have studied the most general case . we find , however , that the pattern by which the extended states disappear as disorder strength increases is rather similar to that of the special case studied earlier . we also explicitly discuss how to interpolate the numerical results on a lattice to the continuum limit properly . the rest of the paper is organized in the following way . in section [ model ] we introduce the model we study and the numerical approach we use to identify extended current carrying states , namely the calculation of a topological quantum number called chern number . in section [ numer ] we present our numerical results for various cases that we have studied . in section [ sum ] we summarize our results , and discuss their relations with previous theoretical and experimental studies . we study numerically non - interacting electrons moving on a lattice with a uniform magnetic field and random onsite potential , described by the following tight binding hamiltonian : @xmath8 where @xmath9 is the fermion operator on lattice site @xmath10 . the first term represents hopping or kinetic energy of the electrons , and the hopping phases @xmath11 are determined by the uniform magnetic field ( up to arbitrary gauge choices ) . we consider the case of @xmath12 flux quantum per square , where @xmath13 and @xmath14 are primitive integers . in this case , the original tight binding band ( for zero field ) is split into @xmath14 landau subbands in the absence of the random potential . the second term represents a random onsite potential . we take @xmath15 to be uniformly distributed between @xmath16 and @xmath17 , measured in units of the magnitude of the nearest neighbor hopping matrix element , which is set to be 1 throughout this paper . the special case of the model ( [ ham ] ) with near neighbor hopping only and @xmath18 has been studied in some detail before.@xcite it is found in this case that for weak randomness , the extended states in most landau subbands form a critical energy with total chern number@xcite ( tcn ) @xmath19 , while those in the central subband ( odd @xmath14 ) or two central subbands ( even @xmath14 ) split into two critical energies with negative tcns . the total chern number of the entire band is zero , as the filled band has zero hall conductance . increasing randomness strength merges critical energies with tcn @xmath19 and those with negative tcns , with those closer to the band center disappearing at lower randomness.@xcite simultaneously , it is found that the critical energies near the band edge float away from the centers of their corresponding landau subbands towards the center of the whole band,@xcite consistent with the floating up " picture@xcite in continuum systems where there is no critical energy with negative tcn . such behavior is understood heuristically as due to an attraction " between states or critical energies with chern numbers of opposite signs.@xcite this special case discussed above has the very unique feature that in the absence of randomness all subbands carry chern number @xmath19 except for the central subband(s ) . in experiments however , the field strength is varied continuously . also restricting the model to near neighbor hopping only endows the model with particle - hole symmetry , which is also absent in real systems@xcite . for general field strength commensurate with the lattice , i.e. , @xmath12 flux quantum per plaquette , the chern numbers of subbands follow some nontrivial pattern and subbands carrying chern numbers of opposite signs intertwine with each other . it is unclear how the extended states behave upon introducing randomness in this case . we address this issue numerically in the present paper , and without losing generality , we assume @xmath20 . to study the localization property of this model , we use the chern number approach,@xcite which is briefly reviewed below . the hall conductance of an individual eigenstate @xmath21 can be obtained easily using the kubo formula : @xmath22 where @xmath23 is the area of the system , @xmath24 and @xmath25 are the velocity operators in the @xmath26 and @xmath27 directions respectively . for a finite system with the geometry of a parallelogram with periodic boundary conditions ( torus geometry ) , @xmath28 depends on the two boundary condition phases @xmath29 and @xmath30 . as shown by niu _ _ , the boundary condition averaged hall conductance takes the form@xcite @xmath31 where @xmath32 is an integer called the chern number of the state @xmath21 . states with nonzero chern numbers carry hall current and are necessarily extended states@xcite . thus by numerically diagonalizing the hamiltonian on a grid of @xmath29 and @xmath30 , and calculating the chern numbers by converting the integral in ( 2 ) to a sum over grid points , we are able to identify extended states unambiguously . depending on system size , the number of grids used in numerical calculations range from @xmath33 to @xmath34 . in this section we present our numerical results . we start by discussing the case when there is only nearest neighbor hopping , and no disorder . in this case we have an exact particle - hole symmetry , and the quantized hall conductance of each of the @xmath14 landau subbands ( which is also the chern number of this subband ) is known exactly . for the @xmath35th subband , the quantized hall conductance ( in unit of @xmath36 ) @xmath37 when it is occupied _ together with all the subbands below it _ , satisfies the diophantine equation@xcite @xmath38 where @xmath39 is an integer that gives rise to the integer @xmath37 with the smallest magnitude satisfying eq . ( [ di]).@xcite the quantized hall conductance of the @xmath35th subband itself , or its chern number , is therefore @xmath40 in the following we show that the solutions of eq . ( [ di ] ) ( together with the condition below it ) has a very non - trivial pattern in it . we start by discussing the case @xmath41 , where @xmath42 is an integer . in this case ( [ di ] ) reduces to @xmath43 for @xmath44 , the solution with @xmath45 having smallest magnitude is realized by choosing @xmath46 , and @xmath47 therefore for @xmath44 , @xmath48 implying that in the lower half of the entire band , _ subbands form groups of @xmath13 subbands with total chern number 1_. from particle - hole symmetry we know that the same happens in the upper half of the entire band as well ; and the remaining @xmath49 subbands at the center of the entire band therefore form a special group carrying negative total chern numbers , so that the total chern number for the entire band is zero . the chern numbers of individual subbands , however , depends sensitively on @xmath13 and @xmath14 , and oscillates between positive and negative integers . therefore in the absence of disorder , the quantized hall conductance does not have a simple monotonous dependence on the filling factor , even if the fermi energy is far away from the band center . also in this case it is not clear how current carrying states move and merge upon introducing disorder , based on the heuristic argument that states carrying chern numbers with opposite signs attract each other".@xcite in the following we show that as randomness is introduced , it first merges the current carrying states _ in the same group_. more specifically , current carrying states in groups away from the band center merge together and form critical energies carrying total chern number @xmath19 , while those in the group at the band center form two critical energies carrying equal amount of negative chern numbers . this is because the energy gaps separating subbands within the same group are much smaller than those separating neighboring groups , therefore the density of states as well as current carrying states in the same group merge together first . at this point the configuration of current carrying critical energies become the same as that of the cases with @xmath18 : there are critical energies with tcn @xmath19 away from the band center , and two critical energies carrying negative tcn near the band center . further increasing the randomness strength will merge the remaining critical energies , with the ones closer to the band center disappearing earlier.@xcite in the next two subsections we present numerical results for cases we have studied , to demonstrate the above observations . it should be borne in mind that the numerical results presented are for finite sized lattices , which may differ somewhat from the results of lattices in the thermodynamic limit ( e.g. , some critical energies which may be disctinct in the latter case may not be resolved for our sizes ) . however , we wish to emphasize that we have taken care to make sure that all the relevant splittings are observable for our sizes , except possibly a small splitting of the central band in the particle - hole symmetric case . we first present our results for cases where there is nearest neighbor hopping only . in these cases we have particle - hole symmetry even in the presence of random potential , _ after _ physical quantities are averaged over all possible disorder configurations . the number of different disorder configurations ( or samples " ) we explore numerically range from 200 to 800 . the special case of @xmath50 has already been studied by tan,@xcite the results obtained by him agree with our general observations above . _ @xmath51_. in this case we have 7 subbands in the absence of disorder , and the chern numbers of the subbands are @xmath52 , obtained from solving eq . ( [ di ] ) with the condition below it , with ones in the same parentheses belong to the same group . in fig.[37 ] we show how the disorder averaged total density of states ( @xmath53 ) , as well as density of current carrying states ( @xmath54 ) , evolve as the randomness strength @xmath17 increases , for systems of square geometry and size @xmath55 . we see for weak randomness of @xmath56 , the tails of @xmath53 of different subbands in the same group have just started to overlap , and there are still well defined peaks of @xmath54 for each individual subband . in the thermodynamic limit , current carrying extended states are expected to exist only at individual critical energies;@xcite in finite size systems however , @xmath54 has finite width around these critical energies as states with localization length bigger than the system size appear extended . as @xmath17 is increased to @xmath57 , subbands in the same group have merged together , so have @xmath54 , while there are still gaps separating different groups . further increasing @xmath17 to @xmath58 , @xmath53 of the 3 different group start to merge ; the current carrying states in the central group ( which consists of only a single subband in this case ) split into two peaks ( that become sharp critical energies each carrying tcn -1 in the thermodynamic limit ) , moving toward the band edges to meet the current carrying states in the side groups , which form two critical energies each carrying tcn + 1 in the thermodynamic limit . at this point , as far as the behavior of current carrying states is concerned , the situation is identical to that of the case with @xmath59 which has been studied closely earlier,@xcite even though initially we have 7 subbands in this case . further increasing @xmath17 to @xmath60 , the critical energies carrying opposite tcn have merged and disappeared , and there are no current carrying states anywhere in the thermodynamic limit ; the system is in an insulating state no matter where the fermi energy is . in a finite size system we see very low @xmath54 which is expected to go to zero rapidly as system size increases . _ @xmath61_. in this case we again have 7 subbands in the absence of disorder , and the chern numbers of the subbands are @xmath62 , again with the ones in the same parentheses belong to the same group . we again study systems with size @xmath55 , and show in fig.[27 ] how @xmath53 and @xmath54 evolve with @xmath17 in this case . for @xmath56 , the subbands in the two side groups have already merged together , while the tails of @xmath53 of the central group ( with 3 subbands ) barely start to overlap . increasing @xmath17 to @xmath57 , the 3 subbands in the central group now have significant overlap , while the gaps separating different groups remain . as in the previous case , further increasing @xmath17 to @xmath58 , @xmath53 of the 3 different group start to merge , and he current carrying states in the central group split into two peaks , moving toward the band edges to meet the current carrying states in the side groups . further increasing @xmath17 to @xmath60 , all critical energies have merged and disappeared , and we see a very low @xmath54 . in this case we have 8 subbands in the absence of disorder , and the chern numbers of the subbands are @xmath64.@xcite we have studied systems with size @xmath65 . the situation is quite similar to the previous two cases . for @xmath66 , @xmath53 of the subbands in the same group have just started to overlap . as @xmath17 is increased to @xmath57 , subbands in the same group have merged together , so have @xmath54 , while there are still gaps separating different groups . further increasing @xmath17 to @xmath67 , @xmath53 of the 3 different group start to merge ; the current carrying states in the central group split into two peaks , moving toward the band edges to meet the current carrying states in the side groups , which form two critical energies each carrying tcn + 1 in the thermodynamic limit . further increasing @xmath17 to @xmath68 , the critical energies carrying opposite tcn have just merged and disappeared . in this case we have 9 subbands in the absence of disorder , and the chern numbers of the subbands are @xmath70 . the dependence of @xmath53 and @xmath54 on @xmath17 for systems of size @xmath71 are shown in fig . [ 29 ] . for @xmath56 , @xmath53 and @xmath54 of groups 1 and 5 have fully merged already , and those in groups 2 and 4 have started to merge , although we still have separate peaks there . gaps between groups are quite pronounced at this point . increasing @xmath17 to 1.5 , the subbands in groups 2 and 4 have also fully merged ; and they have also started to merge with the central group ( consisting of only 1 subband ) . at this point the situation has become identical to that of @xmath72 : we have 4 critical energies each carrying tcn + 1 away from the band center , while the current carrying states at band center has tcn -4 . further increasing @xmath17 to 2.5 , we find the current carrying states from the central group have split into two critical energies , and they in turn have already merged with the current carrying states from groups 2 and 4 ; there remain 4 critical energies as indicated by the 4 peaks in @xmath54 ; 2 of them carry tcn + 1 , from the edge groups ; the other 2 closer to the band center carry tcn -1 . all the remaining critical energies have merged and disappeared as @xmath17 is increased to 3.5 . in this subsection we present data for cases where there is next nearest hopping , with the magnitude of the hopping matrix element @xmath73 , in unit of the nearest neighbor hopping . in the presence of @xmath73 , the particle - hole symmetry present in the model with nearest neighbor hopping only is broken . although eq . ( [ di ] ) is still satisfied in this case , generically the magnitude of @xmath37 is not necessarily minimized ( by choosing the appropriate @xmath39 ) . however if @xmath73 is small enough , and if the subband gaps are not closed as @xmath73 is increased from 0 , the chern number of each subband remains the same as that of @xmath74 . in this case we have 3 subbands . the hall conductance ( chern number ) of each subband is + 1 , + 1 , and -2 . this is _ different _ from the case with @xmath74 , where the chern numbers are + 1 , -2 , + 1 . we find the next nearest neighbor hopping term breaks the particle - hole symmetry , and moves the negative chern numbers to higher subbands . in fig . [ 13 ] we show how @xmath53 and @xmath54 for systems of size @xmath76 evolve with @xmath17 . for weak randomness @xmath56 , the gaps separating different subbands are still finite . increasing @xmath17 to 2.0 , the central and upper subbands have merged , and the current carrying states in both subbands have also merged to form one critical energy represented by a single peak , carrying total chern number -1 . further increasing @xmath17 to 4.0 , the lower subband also starts to merge with the other subbands , and the two remaining critical energies carrying chern numbers + 1 and -1 respectively start to move together . when @xmath17 is increased to 6.0 , all the subbands have merged together completely , and the two critical energies have also merged and disappeared . this is a somewhat special case , because with @xmath74 , the system actually respects time - reversal symmetry ( since the nearest neighbor hopping matrix elements can be chosen to be real ) , and the hall conductances ( and chern numbers ) are zero everywhere . related to this is the fact that the two subbands are degenerate at two points in the brillouin zone in this case . in the presence of a nonzero @xmath73 , a gap separating the two subbands is opened up , and the chern numbers of the lower and upper subbands become 1 and -1 respectively . in fig . [ 12 ] we show results of numerical studies when random potential is added to the system , for systems with size @xmath65 . previously ludwig _ et al._@xcite studied this case analytically . we see when @xmath78 , there is still a gap separating the two subbands , and the current carrying states form two critical energies represented by the peaks in @xmath54 . increasing @xmath17 to 4.0 , the two subbands have started to merge , and the two critical energies start to move closer together toward the band center . further increasing @xmath17 to 5.0 , the two subbands have merged completely , and the two critical energies have almost merged . further increasing @xmath17 to 7.0 , the two critical energies have merged and disappeared , and there are very few current carrying states left in such a finite size system . in this paper we have studied numerically how current carrying states disappear , and the system undergoes phase transitions from a quantum hall state to an insulating state as disorder strength increases , in tight binding lattice models with non - interacting electrons . in cases with nearest neighbor hopping only ( with particle - hole symmetry ) , we find the following generic pattern : subbands away from the center of the band form groups with @xmath13 subbands , and the total chern number of such groups is @xmath19 ; while subbands near the band center for a special group with @xmath49 subbands . in the presence of relatively weak randomness , groups of critical energies away from the band center merge together and form critical energies carrying hall conductance ( or total chern number ) 1 ; the current carrying states near the band center form ( within our resolution ) two critical energies carrying negative chern numbers . further increasing randomness strength , the critical energies carrying negative chern numbers move toward the band edge and merge with the critical energies carrying total chern number 1 , and eventually all critical energies and current carrying states disappear in the thermodynamic limit , and the system becomes insulating no matter where the fermi energy is . adding a next nearest neighbor hopping term breaks the particle - hole symmetry . we find this tends to push subbands carrying negative chern numbers to higher energies . however as long as the fermi energy is near the bottom of the band , the way critical energies carrying positive chern numbers are killed is similar : they are killed by critical energies carrying negative chern numbers coming down from high energy.@xcite our results suggest that the non - trivial structure in the dependence of hall conductance on electron density in the presence of a periodic potential is quite vulnerable against randomness ; in order to see such structures experimentally,@xcite one must control the amount of randomness carefully so that subbands in the same group ( which are separated by very small gaps ) are not merged together by randomness . as the randomness becomes strong enough to merge subbands in the same group , there remain only critical energies carrying hall conductance 1 except near the band center . thus the situation is essentially the same as without periodic potential , i.e. , there is no oscillation in the dependence of the hall conductance on electron density . in the one - band tight binding model studied here and elsewhere , all current carrying extended states disappear at finite randomness strength for a fixed strength of magnetic field . based on this observation , xie _ et al._@xcite proposed that in the _ continuum _ system , instead of floating to infinite energy in the limit @xmath1 , all extended states disappear at _ finite _ @xmath0 for a given disorder strength . in the following we show that extrapolating the lattice model to the continuum limit is quite subtle , and one needs to be extremely careful in interpreting the implications of results of lattice model on the continuum system . for simplicity we consider the case @xmath18 ; extension to cases with @xmath79 in the following discussion is straightforward.@xcite without randomness , there are two energy scales in the lattice model for a given strength of magnetic field , namely , the bandwidth which is of order @xmath35 ( the hopping matrix element ) , and the landau level spacing @xmath80 , where @xmath81 is the effective mass at the bottom of the band . there are also two length scales , the lattice spacing @xmath82 and the magnetic length @xmath83 . in the continuum system however , there is only one energy scale , namely the landau level spacing @xmath84 , and one length scale , the magnetic length @xmath85 ; the band width is infinite , and the lattice spacing is zero . thus to reach the continuum limit in the tight binding lattice model , one must send both @xmath35 and @xmath14 to infinity , with the ratio @xmath86 remain a constant ; in the mean time @xmath82 should be sent to zero in such a way that @xmath87 remains a constant . the random potential introduces another energy scale.@xcite naively , the new scale is @xmath17 ; what has been shown in lattice models is that the critical @xmath88 above which no extended states exist is finite _ in units of t_,@xcite and @xmath88 goes to zero as @xmath14 goes to infinity , again _ in units of t_. in the continuum model however , the real new energy scale introduced by the random potential is the amount of broadening of a landau level , @xmath89 . the important question in the continuum model is _ in units of @xmath84 _ , how much @xmath89 does one need to kill all the current carrying extended states , or what is the ratio @xmath90 ? if the answer is finite , then for a given disorder strength , one needs a finite critical magnetic field in the continuum for extended states to exist , as suggested in refs . ; while if the answer is infinite , then the extended state must exist no matter how small @xmath0 is , and float to infinite energy only in the limit @xmath1 , as originally proposed by khmelnitskii@xcite and laughlin.@xcite in the lattice model with large @xmath14 , we have @xmath91 . this is because the linear size of the wave function in a given landau level is @xmath85 , which means the number of lattice sites it occupies is proportional to @xmath92 ; since the random potential on different sites are uncorrelated , the shift on the energy is typically of order @xmath93 , the magnitude of the _ averaged _ potential in that area . therefore we find @xmath94 let us assume @xmath95 in the limit @xmath96 . since @xmath89 must be at least comparable to @xmath84 for current carrying states to disappear , from eq . ( [ ratio ] ) we must have @xmath97 . if @xmath98 , the ratio in eq . ( [ ratio ] ) is finite as @xmath96 ; while if @xmath99 , or @xmath98 but there is logarithmic corrections to eq . ( [ asp ] ) , the ratio in eq . ( [ ratio ] ) becomes infinite as @xmath96 ; the situation becomes qualitatively different . it was shown in ref . that the data for @xmath100 may be fit reasonably well for @xmath98 . since the crucial question is whether @xmath101 is _ exactly _ @xmath102 or not , we feel it is probably difficult to resolve this issue numerically , and analytic input is needed here . the authors are particularly grateful to allan macdonald and shivaji sondhi for important conversations . they have also benefited from useful discussions and correspondences with mike hilke , danny shahar , dong - ning sheng , s .- h . song , david thouless , dan tsui , zheng - yu weng , and xin - cheng xie . this work was supported by nsf grant dmr-9400362 ( at princeton ) , and the sherman fairchild foundation ( at caltech ) . r. n. b. acknowledges the hospitality of the aspen center for physics while this manuscript was being written . for reviews , see _ the quantum hall effect _ , edited by r. e. prange and s. m. girvin ( springer - verlag , new york , 1990 ) ; _ perspectives in quantum hall effect _ , edited by s. das sarma and a. pinczuk ( wiley , new york , 1997 ) . the critical energies are exactly at the centers of landau bands only for electron - hole ( e - h ) symmetric models . in real semiconductor systems , this is only approximately true , with the e - h symmetry broken both by ( i ) a non - symmetric random potential , and ( ii ) mixing between different landau levels . there is experimental evidence ( s. v. kravchenko _ et al . b * 50 * , 8039 ( 1994 ) ) that a metallic phase may be stabilized by strong electron - electron interactions at @xmath103 . we do not consider this possibility here . strictly speaking in this case ( and all other cases with even @xmath14 ) , the chern numbers of the two central bands are not well defined , as there is no gap separating them ; only the sum of their chern numbers are well defined . however a gap may be opened by introducing a very small next neighbor hopping term to make the chern numbers well defined . for early experimental attempts in studying the iqhe in the presence of a periodic potential , see r. r. gerhardts , d. weiss , and k. v. klitzing , phys . rev . lett . * 62 * , 1173 ( 1989 ) ; r. w. winkler , j. p. kotthaus , and k. ploog , phys . lett . * 62 * , 1177 ( 1989 ) . for the uncorrelated random onsite potential studied here ( which maps onto gaussian white noise potential with zero correlation length in the continuum limit ) , no new length scale is introduced , although it would be interesting to study random potential with finite correlation length as well .
we report results of numerical studies of the integer quantum hall effect in a tight binding model on a two - dimensional square lattice with non - interacting electrons , in the presence of a random potential as well as a uniform magnetic field applied perpendicular to the lattice . we consider field magnitudes such that the area per flux quantum is commensurate with the lattice structure . topological properties of the single electron wave functions are used to identify current carrying states that are responsible for the quantized hall conductance . we study the interplay between the magnetic field and the disorder , and find a universal pattern with which the current carrying states are destroyed by increasing disorder strength , and the system driven into an insulating state . we also discuss how to interpolate results of lattice models to the continuum limit . the relationship to previous theoretical and experimental studies of quantum hall - insulator transitions in strongly disordered systems at low magnetic fields is discussed .
it is well known that superconductivity and ferromagnetism are two competing orders . the coexistence of singlet superconductivity and ferromagnetism is basically impossible in the same compound but may be easily achieved in artificially fabricated superconductor / ferromagnet ( s / f ) hybrid structures . in this case , the coexistence of the two orders is due to the proximity effect @xcite . the main manifestation of the proximity effect in s / f structures is the damped oscillatory behavior of superconducting correlations in the f layer . two characteristic lengths of the decay and the oscillations are , respectively , @xmath0 and @xmath1 . therefore in s / f heterostructures there is a unique possibility to study the properties of superconducting electrons under the influence of the exchange field in the ferromagnet . recent progress in the preparation of the high quality s / f layered structures permitted to experimentally observe many striking phenomena that are quite interesting for applications such as a non - monotonic dependence of their critical temperature and oscillations of critical current in s / f / s josephson junctions as a function of the f layer thickness ( see @xcite and references therein ) . it is possible to fabricate josephson @xmath2 junctions with a @xmath2-phase difference in the ground state , which are good candidates for elements in superconducting logic circuits @xcite . sifs junctions , i.e. s / f / s trilayers with one transparent interface and one tunnel barrier between s and f layers , represent an interesting case of @xmath2 junctions for applications where active josephson junctions are required . the sifs structure offers the freedom to tune the critical current density over a wide range and at the same time to realize high values of the product of the junction critical current @xmath3 and its normal state resistance @xmath4 @xcite . in addition , nb based tunnel junctions are usually underdamped , which is desired for many applications . sifs junctions are also interesting from the fundamental point of view since they provide a convenient model system for a comparative study between @xmath5-@xmath2 transitions observed from the critical current and from the density of states ( dos ) @xcite . the purpose of this work is to provide a quantitative model describing the dos in sifs junctions and to calculate current - voltage characteristics of a sifs junction . the latter may be used for estimation of the dissipation in sifs - based qubit systems . the model of an s / f / s junction we are going to study is depicted in fig . [ sifs ] and consists of a ferromagnetic layer of thickness @xmath6 and two thick superconducting electrodes along the @xmath7 direction . the left superconductor electrode is voltage - biased . left and right superconductor / ferromagnet interfaces are characterized by the dimensionless parameters @xmath8 and @xmath9 , respectively , where @xmath10 , @xmath11 are the resistances of the left and right s / f interfaces , respectively , @xmath12 is the conductivity of the f layer , @xmath13 , @xmath14 is the diffusion coefficient in the ferromagnetic metal and @xmath15 is the critical temperature of the superconductor ( we assume @xmath16 ) . we also assume that the s / f interfaces are not magnetically active . we will consider the diffusive limit , in which the elastic scattering length @xmath17 is much smaller than the characteristic decay length @xmath18 . = 7.5 cm in this paper we concentrate on the case of a sifs tunnel josephson junction , when @xmath19 ( tunnel barrier ) and @xmath20 ( fully transparent interface ; however in our numerical calculations we use finite but very small @xmath21 ) . in this case the left s layer and the right fs bilayer in fig . [ sifs ] are decoupled . therefore we can calculate the quasi - particle current through a sifs junction using the standard formula of the tunnel theory @xcite , @xmath22,\ ] ] where @xmath23 is the bcs density of states [ @xmath24 is the heaviside step function ] and @xmath25 is the density of states in the fs bilayer at the free boundary of the ferromagnet ( @xmath26 ) . both @xmath27 and @xmath25 are normalized to their values in the normal state ; @xmath28^{-1}$ ] is the fermi function . to obtain @xmath25 we should solve the usadel equations in the ferromagnetic layer of the fs bilayer . using the @xmath29-parameterizations of the normal and anomalous green functions , @xmath30 , @xmath31 , we can write the usadel equations in the f layer as @xcite @xmath32 where a positive ( negative ) sign in front of @xmath33 corresponds to the spin up state @xmath34 ( spin down state @xmath35 ) , @xmath36 are the matsubara frequencies , @xmath33 is the exchange field in the ferromagnet , and the parameter @xmath37 is the spin - flip scattering time . we consider a ferromagnet with strong uniaxial anisotropy , in which case the magnetic scattering does not couple the spin up and spin down electron populations . in the s layer the usadel equations take the form ( where we omit subscripts ` @xmath38 ' because equations in superconductor look identically for spin up and spin down electron states ) , @xmath39 they should be completed with the self - consistency equation , @xmath40 here @xmath41 is the diffusion coefficient in the superconductor and @xmath42 is the superconducting pair potential . since we calculate the dos at the free boundary of the ferromagnet in the fs bilayer , we need to set to zero the @xmath43 derivative at the left s / f interface , @xmath44 at the right f / s interface the boundary conditions are given by the relations @xcite , [ kl ] @xmath45 where @xmath46 , @xmath47 is the conductivity of the s layer and @xmath48 . the parameter @xmath49 determines the strength of suppression of superconductivity in the right s lead near the interface compared to the bulk : no suppression occurs for @xmath50 , while strong suppression takes place for @xmath51 . in our numerical calculations we will assume small @xmath52 . to complete the boundary problem we also set a boundary condition at @xmath53 , @xmath54 = 7.5 cm = 7.5 cm to calculate the dos @xmath25 on the free f layer boundary ( @xmath26 ) we use the self - consistent two - step iterative procedure @xcite . in the first step we calculate the pair potential s coordinate dependence @xmath42 using the self - consistency equation in the s layer , eq . ( [ delta ] ) . then , by proceeding to the analytical continuation in eqs . ( [ usadel ] ) , , , , of the quasi - particle energy @xmath55 and using the @xmath42 dependence obtained in the previous step , we find the green functions by repeating the iterations until convergency is reached . we define the full dos @xmath25 and the spin resolved dos @xmath56 , normalized to the dos in the normal state , as [ dos ] @xmath57/2 , \label{dos_full } \\ n_{\uparrow(\downarrow)}(e ) & = \operatorname{\mathop{\mathrm{re}}}\left[\cos\theta_{b \uparrow(\downarrow)}(i\omega \rightarrow e + i0)\right ] , \label{dos_spin}\end{aligned}\ ] ] where @xmath58 is the boundary value of @xmath43 at @xmath59 . in case of a long f layer ( @xmath60 ) it is also possible to obtain an analytical expression for the dos at the free boundary of the ferromagnet @xcite . in fig . [ dosa0 ] and [ dosa1 ] we plot the dos energy dependence for different @xmath61 . fig . [ dosa0 ] corresponds to the case of the absence of spin - flip scattering ( infinite @xmath37 ) and fig . [ dosa1 ] to the case of @xmath62 . in both figures we plot the dos for two chosen values of exchange field , @xmath63 for plots ( a)-(d ) and @xmath64 for plots ( e)-(h ) . in fig . [ dosa0 ] we see that for any chosen @xmath33 at small @xmath61 the full dos turns to zero inside a mini - gap , which vanishes with the increase of @xmath61 . then the dos at the fermi energy @xmath65 rapidly increases to the values larger than unity and with further increase of @xmath61 it oscillates around unity while it s absolute value exponentially approaches unity . in the case of long enough ferromagnets we can observe dos peaks at @xmath66 [ we notice that @xmath67 . also for @xmath68 @xcite we can use an analytical approximation for @xmath25 @xcite . for smaller @xmath61 it is incorrect to use this approximation . it is shown by the dashed red line and is in rather good agreement with the numerical solution . in the numerical curves the peaks at @xmath66 are smeared because we used finite @xmath69 for the transparent f / s interface at @xmath70 . in fig . [ dosa1 ] we observe similar tendencies . however for smaller @xmath37 the mini - gap closes at smaller @xmath61 , the period of the dos oscillations at the fermi energy increases and the damped exponential decay occurs faster . also the dos peak at @xmath66 is smeared . = 7.5 cm = 7.5 cm we calculate the current - voltage characteristics ( cvc ) of a sifs junction at @xmath71 using eq . ( [ i(v ) ] ) . plots for different parameters are shown in figs . [ iva0 ] and [ iva1 ] . for comparison we also present the cvc of a sins tunnel junction , i.e. a junction with a normal metal interlayer instead of a ferromagnet . we can see that for a certain range of parameters the cvc of a sifs junction exhibits an interesting @xmath72-like feature [ we can observe it for @xmath73 in fig . [ iva0 ] ( b ) and for @xmath63 in fig . [ iva0 ] ( c ) ] . it corresponds to the case of a large subgap dos in fig . [ dosa0 ] ( c),(f ) . for finite spin - flip scattering the features of cvc are smeared . at large @xmath61 they totally disappear . a more detailed analysis of cvc of sifs junctions will be presented elsewhere @xcite . to summarize , we calculated the quasi - particle dos in the f layer of a sifs junction in close vicinity of the tunnel barrier and use it to obtain the current - voltage characteristics of a sifs junction . the developed formalism may be used for estimations of dissipation in sifs - based qubits , which will be done elsewhere . e. terzioglu and m. r. beasley , ieee trans . appl.supercond . * 8 * , 48 ( 1998 ) ; l. b. ioffe , v. b. geshkenbein , m. v. feigelman , a. l. fauchre , and g. blatter , nature , * 398 * , 679 ( 1999 ) ; g. blatter , v. b. geshkenbein , l. b. ioffe , phys . b * 63 * , 174511 ( 2001 ) ; a. v. ustinov and v. k. kaplunenko , j. appl . phys . * 94 * , 5405 ( 2003 ) ; t. ortlepp , a. ariando , o. mielke , c. j. m. verwijs , k. f. k. foo , h. rogalla , f. h. uhlmann , h. hilgenkamp , science , * 312 * , 1495 ( 2006 ) .
we investigate superconductor / insulator / ferromagnet / superconductor ( sifs ) tunnel josephson junctions in the dirty limit , using the quasiclassical theory . we consider the case of a strong tunnel barrier such that the left s layer and the right fs bilayer are decoupled . we calculate quantitatively the density of states ( dos ) in the fs bilayer for arbitrary length of the ferromagnetic layer , using a self - consistent numerical method . we compare these results with a known analytical dos approximation , which is valid when the ferromagnetic layer is long enough . finally we calculate quantitatively the current - voltage characteristics of a sifs junction . , , , , sifs junction , density of states , dissipative current 74.45.+c , 74.50.+r , 74.78.fk , 75.30.et
the emergence of drug resistance represents one of the major clinical challenges of the current century @xcite . microbial pathogens quickly acquire resistance to new antibiotics @xcite , while solid tumors often regrow after treatment because of resistance mutations that arise during tumor growth @xcite . in addition to genomic studies examining the molecular causes of resistance @xcite , the dynamics of drug resistance evolution has recently attracted wide interest @xcite , with the dual goal of understanding the emergence of resistance and developing novel strategies to prevent or control its spread @xcite . next - generation sequencing and high - throughput experimental techniques enable the quantitative study of resistance evolution but require the development of new theories to appropriately interpret experimental results @xcite . in many realistic systems , an evolving population interacts with its surroundings and exhibits a well - defined spatial structure ( for instance , in tumors and biofilms @xcite ) . it has recently been shown that this spatial structure can strongly influence the subclonal structure and the adaptation of spatially expanding populations , both from _ de novo _ and pre - existing mutations @xcite . likewise , spatial @xcite or temporal @xcite gradients in antibiotic concentration can enable populations to reach a higher degree of resistance than in homogeneous drug concentrations , at least in part because they enable the slow accumulation of multiple mutations , each conferring a small amount of resistance . the presence of spatial drug gradients is well documented both in the outside environment @xcite as well as within biofilms @xcite and the human body @xcite , and it has been hypothesized that the presence of spatial heterogeneities may facilitate the emergence of drug - resistant phenotypes @xcite . in a microfluidic experiment , a spatial gradient indeed gave rise to a higher rate of adaptation of bacterial populations @xcite . similarly , microbes growing on soft agar plates with gradually increasing antibiotic concentrations were able to rapidly evolve resistance to high levels of antibiotics , while sudden jumps to unsustainably high concentrations dramatically slowed down adaptation @xcite . these findings raise the theoretical question of how to predict the rate of emergence of drug resistance in the presence of spatial gradients . a number of recent theoretical studies have investigated how gradients speed up the evolution of drug resistance @xcite . each study made critical assumptions about the nature of the gradients : greulich et al . @xcite considered a population adapting to a smooth gradient , which gradually lowers the growth rate of susceptible individuals . hermsen et al . @xcite studied resistance evolution in a series of sharp step - like increases in concentration , where a novel resistance mutations was necessary for survival in the next step ( the `` staircase '' model ) ; hermsen @xcite later proposed a generalization of the staircase model to continuous gradients . these previous studies focused on the speed of adaptation , i.e. , how quickly the population evolves to tolerate high concentrations of antibiotics . in the context of the emergence of drug resistance , this observable alone ignores a crucial reality of antibiotic treatment : efficient drug treatment first and foremost aims to kill as many bacteria as possible , while limiting the rise of resistance mutation . how this apparent trade - off can be optimized to prevent the evolution of drug resistance has so far been unexplored . moreover , many realistic growth scenarios of bacterial populations may exhibit a directed flow that drives them up or down the gradient . examples include the gut , arteries , and urethra in the human body @xcite , but also flows in aquatic environments , like ocean and river currents @xcite , or flow in pipes and catheters @xcite . the effect of convection on the evolution of drug resistance remains unknown . here , we present simulations , rationalized by a comprehensive analytical framework , of populations evolving resistance in a variety of spatial antibiotic concentration gradients and under the influence of convection . we measure the establishment probability of resistant mutants arising in a region occupied by susceptible wild type , and the drug - induced death rate of the wild type , and show how bacterial diffusion , antibiotic gradient steepness , and convection interact to affect the treatment efficiency . we find that shallow gradients and convection into the antibiotic promote wild - type death , at the cost of increasing the establishment probability of resistance mutations . conversely , populations in steep gradients and subject to convection away from the antibiotics are less susceptible to drug - induced wild - type death but also produce fewer resistance mutants . the treatment efficiency , which quantifies the inherent trade - off between adaptation and death , is strongly modulated by gradient steepness and convection . treatment efficiency is found to strongly depend on convection away from the antibiotic , which increases it in shallow gradients and decreases it in steep gradients . we perform individual - based , stepping stone simulations where both wild - type and mutants are modeled explicitly . the population is divided into demes with carrying capacity @xmath0 on a one - dimensional lattice . wild - type and mutants replicate at a rate @xmath1 and migrate at rate @xmath2 independently on their position . wild - type in deme @xmath3 die at a rate @xmath4 $ ] . since we assume throughout that the antibiotic leaves the mutants unaffected , mutants do not die in our simulations . analogously to a gillespie algorithm , in each simulation step , a birth , death or migration event is performed according to its relative rate @xcite , as follows . * * birth . * birth events occur at a total rate equal to @xmath5 , where @xmath6 is the total number of individuals in deme @xmath3 . for each birth event , a source individual is selected at random and replicated into a random target site between 1 and k within the same deme . because the target site can either already be filled with an individual or be empty , this move effectively translates into logistic growth within the deme . * * in our model , only the wild type can die , thus deaths have a total rate corresponding to @xmath7 , where @xmath8 represents the number of wild type individuals in deme @xmath3 . to perform a death event , first , a deme @xmath3 is picked proportionally to its relative death rate @xmath9 , and then , a random wild type within the same deme selected to be removed . * * migration . * migrations are performed at a rate @xmath10 by picking a random individual and swapping it with a randomly selected target site from one of the two neighboring demes . as in the case of birth events , the target site can either correspond to an individual , or to an empty site . * * time step . * time is tracked by sampling a time interval @xmath11 from an exponential distribution with rate @xmath12 $ ] , as in a standard gillespie algorithm . the total elapsed time is the sum of the sampled time intervals . * * convection . * convection ( with convection speed @xmath13 ) is implemented by shifting the simulation box by one deme towards or away from the wild type population , depending on whether convection is negative or positive , respectively . the shift is performed when the time since the last shift is greater than @xmath14 . for each simulation , we first allow the wild type to reach the steady - state profile @xmath15 . we then introduce one mutant element at position @xmath16 and run the simulation until either all mutants go extinct , or mutants reach the last deme in the simulation box . no further mutations are allowed in the course of the simulations . the probability of fixation @xmath17 is then computed as the proportion of the simulations in which a mutant introduced at @xmath16 reached fixation . all numerical results were obtained by evaluating the differential equations using mathematica s built - in _ ndsolve _ routine with the backwards differentiation ( bdf ) method , with a maximum step size of 1 and a domain size of 1000 . initial conditions were chosen according to the analytical approximations to the steady - state profiles given in the text . this is only done to speed up computation and increase numerical stability ; starting with different initial conditions leads the same final solution . to obtain steady - state profiles , we solve the full time - dependent problems until the solution no longer changes for longer evaluation times . to compute the establishment probability @xmath17 in realistic wild - type population profiles @xmath15 , we first computed the steady - state population density and then used the final profile to compute the ( constant in time ) local death rate for the mutants @xmath18 . the resulting numerical solutions were integrated numerically using mathematica s built - in _ nintegrate _ routine to obtain the total death rate @xmath19 and the establishment score @xmath20 . ) or away from it ( @xmath21 ) . our goal is to analyze the fixation probability of resistance mutation for a given the rate of wild - type killing.,scaledwidth=49.0% ] , defined as the number @xmath19 of drug - induced deaths per generation divided by the establishment score @xmath20 ( see eq . [ eq : q ] ) , for different antibiotic concentration profiles ( shallow gradient , @xmath22 ; intermediate gradient , @xmath23 ; step - like concentration profile ) . convection away from the antibiotic ( positive @xmath13 ) can increase treatment efficiency by an order of magnitude in shallow gradients . ( b ) both @xmath19 and @xmath20 are increased ( in contrast to their ratio @xmath24 ) in shallow gradients and for convection into the antibiotic.,scaledwidth=49.0% ] ( magenta ) and establishment probability @xmath17 ( green ) for three different concentration profiles ( gray background ; ( a ) shallow gradient , @xmath22 ; ( b ) intermediate gradient @xmath23 ; ( c ) step ) and different values of @xmath25 to @xmath26 ( light to dark colors ) , from stepping stone simulations . dotted and dashed line are the local wild - type death rate @xmath27 and the net growth rate @xmath28 . on the right in each panel is the product @xmath29 , which identifies the region where successful mutants arise.,scaledwidth=49.0% ] we simulate a population of wild - type individuals on a lattice of @xmath30 demes , each of size @xmath0 , in a one - dimensional antibiotic gradient ( see the sketch in fig . [ fig : sketch ] and _ methods _ ) . each individual can migrate into a neighboring deme , replicate , and die . the antibiotic gradient sets the death rate of the wild type , giving rise to an effective growth rate @xmath31 . \label{eq : tanh}\ ] ] in the absence of antibiotics , population growth is only limited by a carrying capacity @xmath0 . following the equilibration to the steady - state profile of the wild type , a resistant mutant is inserted into the population at position @xmath16 . the resistant mutant has the same birth rate as the wild type , but it does not suffer from an increased death rate due to the antibiotics . we follow the mutant clone until it either goes extinct or reaches the far end of the simulation box , in which case we consider the mutant _ established_. in treating a bacterial infection , the optimal antibiotic strategy would eliminate as many bacteria as possible , while limiting the emergence of resistance . to quantify the treatment efficacy , we first obtain an average wild - type population profile @xmath15 from independent simulation runs and thus compute the number @xmath19 of drug - induced wild - type deaths per generation , @xmath32 where @xmath33 is the death rate per generation due to antibiotics in deme @xmath34 . for simplicity , we call @xmath19 the total death rate . to quantify the emergence of resistance , we measure the local mutant establishment probability @xmath17 , i.e. , the probability that a mutation that arose at position @xmath16 establishes . since the probability that a mutation occurs in the first place is proportional to the wild - type population density @xmath15 , it follows that successful mutants can only arise where both the wild - type population density and the establishment probability are high ( see fig . [ fig : cofxanduofx ] ) . a measure for how readily new resistant mutants establish is thus given by the product of wild - type population density and the establishment probability , summed over all demes @xmath34 @xcite . we call this measure the establishment score @xmath20 , @xmath35 the establishment score @xmath20 is proportional to the rate @xmath36 of establishment , the rate at which new resistance mutations arise ( at a low mutation rate @xmath37 ) and establish in the population . alternatively , the establishment score @xmath20 can be understood as a measure proportional to the mean establishment probability @xmath38 , i.e. , the probability that a mutation arising anywhere in the population establishes . finally , we define the efficiency @xmath24 of drug treatment via the amount of drug - induced wild - type death @xmath19 before an adaptive mutation establishes , i.e. , by @xmath39 . we assume mutation rates to be low such that no clonal interference occurs . then , we can treat each mutation independently and thus define @xmath40 fig . [ fig : roverbandrandb]a shows how the treatment efficiency @xmath24 changes across gradient steepness and convection into ( @xmath41 ) and away from ( @xmath21 ) the gradient . we find that without convection ( @xmath42 ) , and for convection toward the gradient ( @xmath41 ) , @xmath24 is slightly higher for shallow gradients but always bounded between @xmath43 and @xmath44 . however , the trend changes dramatically for positive convection away from the gradient , which boosts treatment efficiency by a factor of 10 for shallow gradients and decreases it for the step - like antibiotic profile . considering the establishment score @xmath20 and the total death rate @xmath19 separately ( fig . [ fig : roverbandrandb]b ) illustrates the reason behind the qualitative behavior of treatment efficiency . both @xmath20 and @xmath19 show similar tendencies for different gradients and convection speeds : shallow gradients are characterized by larger values of both @xmath20 and @xmath19 than the step - like concentration profile . negative convection ( towards the gradient ) increases both @xmath20 and @xmath19 by up to a factor of 5 , while positive convection leads to their rapid decrease by two orders of magnitude . remarkably , in shallow gradients the establishment score @xmath20 is more sensitive than the total death rate @xmath19 to positive convection . the rapid decrease of @xmath20 explains the increase in the treatment efficiency @xmath45 for shallow gradients . a detailed look at the population density @xmath15 and establishment probability @xmath17 profiles is necessary to understand the values taken by @xmath20 and @xmath19 . these profiles are shown in fig . [ fig : cofxanduofx ] for three different gradients ( shaded areas ) . the population density @xmath15 approximately tracks the net growth rate @xmath28 , while the establishment probability @xmath17 is roughly a mirror image of the population density . in the region where the drug concentration is low and the wild - type population is dense , the mutants have no advantage over the wild type , since they compete for the same resources , and are likely to go extinct due to genetic drift . conversely , in high drug concentration regions , the wild type can not survive and the mutants can grow freely and establish with high probability . convection has a rich variety of effects on both profiles . convection into the antibiotic ( co - flow , @xmath41 ) leads to broader profiles of both the wild type population density @xmath15 and the establishment probability @xmath17 ( fig . [ fig : cofxanduofx ] ) , while convection away ( counter - flow , @xmath21 ) from the antibiotic reduces both the population density and the establishment probability in the antibiotic region . for shallow gradients , we even observe an apparent cut - off in @xmath15 ( and correspondingly in @xmath17 ) , whose location depends on @xmath13 . consequently , co - flow increases total death rate and establishment score , while counter - flow leads to their rapid decrease . in summary , our simulations show that both the population density profile @xmath15 as well as the local establishment probability @xmath17 of resistance mutants are strongly influenced by environmental parameters , in particular , by the steepness of the antibiotic gradient and the strength of convection . as a general rule of thumb , shallow gradients and convection towards increasing drug concentrations increases the establishment score @xmath20 , while steep gradients and convection away from the drug source decrease it . the total death rate @xmath19 follows the same trend . however , subtle differences between @xmath20 and @xmath19 give rise to a qualitative change in behavior of the treatment efficiency @xmath45 , which is increased by positive convection ( counter - flow ) in shallow gradients and decreased in steep ones . to rationalize our results and identify the critical length scales , we now develop a mathematical model that accommodates these processes . consider a population at some density @xmath15 in an antibiotic concentration field , which gives rise to a death rate @xmath27 per generation , i.e. , the drug - induced death rate is measured relative to the maximum growth rate @xmath1 . we define the total ( drug - induced ) death rate @xmath19 per generation through a continuum version of eq . [ eq : b ] as @xmath46 resistance mutations arise at a rate @xmath37 , which we assume to be constant and sufficiently low such that adaptation occurs via the sequential acquisition of resistance mutations , i.e. , we neglect local clonal interference . analogously to eq . [ eq : r ] above , we define the establishment score @xmath47 where @xmath17 represents the probability that a beneficial mutation born at position @xmath16 eventually establishes . the population density @xmath15 can be computed using standard techniques ( like deterministic reaction - diffusion equations ) , and we return to this question later on . the establishment probability @xmath48 , i.e. , the probability for a mutation to survive until time @xmath49 , obeys a nonlinear reaction - diffusion equation @xcite , @xmath50 where @xmath51 is the local birth rate of the mutants and @xmath52 is the net growth rate ( see si section 1 ) . throughout this paper , we will make the assumption that the birth rate of wild - type and mutant is identical and constant , @xmath53 , and that the drug - induced death rate of the resistant mutant is zero , while the wild - type drug - induced death rate @xmath27 ranges from 0 to @xmath1 , i.e. , we assume that the antibiotic only affects the population by increasing the death rate of the wild type . we include diffusion and convection terms to account for the random dispersal and the effects of external flow on the bacteria . it is important at this point to note that in general the selective gradient @xmath28 will intimately depend on various environmental factors , including competition with the wild - type population . in later sections , especially when incorporating convection ( i.e. , @xmath54 ) , we will take this effect into account . for now , we will assume that this function is known for given external conditions in order to gain a better intuition for the system . we will first analyze eq . [ eq : survival - probability0 ] in the case without drift , i.e. @xmath42 , in order to extract the characteristic length scales of the system . we solve the extreme cases of a step - like concentration increase and a very smooth gradient analytically , and then interpolate between the two regimes using numerical solutions to eq . [ eq : survival - probability0 ] . we begin by considering the simplest functional form for a selective gradient a step : @xmath55 where @xmath56 is the heaviside step function . such a sharp gradient could emerge , for instance , at the boundary of different tissues or organs with different affinities to store antibiotics @xcite . the equation for the fixation probability in this case is given by : @xmath57 where we have rescaled by the diffusion length scale : @xmath58 in the appendix , we solve this equation using an analogy from classical mechanics . we have that for @xmath59 @xmath60 implying @xmath61 for large @xmath16 . for @xmath62 , we obtain @xmath63 - 1\right\ } , \label{eq : positive - u}\ ] ] where @xmath64 . [ eq : step-1 ] shows that the fixation probability decays over a characteristic length scale @xmath65 . this length scale can be roughly understood as the typical distance that a mutant individual travels through random dispersal before replicating . having identified the characteristic length scale over which the establishment probability decays in a step - like concentration profile , we now turn towards more realistic gradients . for simplicity , for the remainder of our calculations , we will model a shallow antibiotic gradient again as the sigmoidal function @xmath66 given in eq . [ eq : tanh ] , @xmath67 $ ] , such that the gradient changes over a characteristic distance @xmath68 . intuitively , if @xmath69 , individuals feel sharp , step - like antibiotic transitions within their typical migration distance , and thus , we expect the results of the previous section to remain good approximations . if , conversely , individuals migrate only short distances compared to @xmath70 , i.e. , @xmath71 , they will sample only a small region of the gradient and will not feel the differences in antibiotics concentration . to make this heuristic idea quantitative , we rescale @xmath72 in eq . [ eq : survival - probability0 ] by the new length scale @xmath70 to obtain @xmath73 when @xmath71 , i.e. , @xmath74 , we can neglect the partial derivative in eq . [ eq : alpha1 ] , such that @xmath75 this so - called quasi - static approximation is a straight - forward extrapolation of the well - mixed result and has been used to model establishment in ref . @xcite . for @xmath76 , the step - like solution holds , whereas , for @xmath77 , no analytical solution is available ; in such cases , we solve the equations numerically ( see methods ) . to determine which scenario is more relevant in microbial communities , we estimate a typical @xmath78 . a typical ( non - motile ) bacterial cell may have a diameter of 1@xmath79 , swimming in a medium of viscosity comparable to that of water ( with blood typically only a factor 3 larger than that @xcite ) . then , its diffusion constant is of order 0.1 - 1@xmath80 . the length scale @xmath65 for a typical birth rate of @xmath81 is then between 50 and a few hundred microns . in a microfluidic experiment by zhang et al . @xcite , demonstrating facilitation of adaptation through antibiotic gradients , the length scale on which the drug gradient varied was @xmath82 m so that @xmath83 @xcite . this estimate is , however , very crude : firstly , the diffusion constant of a bacterium may not be due to thermal fluctuations but due to directed motion that only becomes approximately diffusive on long time scales @xcite . in that case , the diffusion constant can be much larger , up to tens or even 100@xmath80 . for a bacterium with a division time of 60 minutes , this leads to a expected typical length @xmath84 . indeed , for motile bacteria such as those used in a recent experimental study by baym et al . @xcite with a reported spreading velocity of 40mm / hr , we find @xmath85 mm . surely , we can not expect homogeneous antibiotic concentration over such long length scales and indeed step - like transitions in antibiotic concentration were assumed in the study . therefore , we expect that concentration gradients of significant sharpness should play a key role in both experimental set - ups and practically relevant systems , and hence the quasi - static approximation in eq . [ eq : quasi - static ] should be employed with caution . as mentioned above , the mutants net growth rate @xmath28 generally depends on several environmental factors , including competition with the wild - type population . in addition , both the establishment score @xmath20 and the total death rate @xmath19 determined in our simulations depend on the wild - type density profile @xmath15 . to provide analytical expressions for these quantities and the corresponding treatment efficiency @xmath24 , we therefore need to accommodate the coupling between wild - type and mutants . in the following , we will employ a simple logistic growth model for the wild - type population and its interaction with the mutant dynamics . we assume that the wild - type population density @xmath86 is described by the following reaction - diffusion equation : @xmath87 where @xmath88 is the local wild type birth rate , @xmath89 is the local antibiotic - induced death rate of the wild type . in our model , wild - type growth is limited by a logistic term @xmath90 , which models a finite amounts of nutrients limiting the population size per deme to the carrying capacity @xmath0 , while the local death rate @xmath89 acts on each individual independently and is therefore unaffected by the local population density . this distinguishes our model from the fisher equation used to describe population spreading @xcite , where also the death rate multiplies the logistic term . our model instead ensures that the steady - state local population density @xmath91 depends explicitly on the local death rate @xmath89 when the death and birth rate profiles change sufficiently slowly in space , @xmath92 our model , like the original fisher equation , ignores the discrete nature of individuals , which has been shown to significantly alter the tip of the population front ( see si section 7 for a detailed discussion ) . nevertheless , we expect good agreement between our model and simulations in terms of the treatment efficiency @xmath24 , whose value is not significantly affected by the details of the population profile at the tip of the wave . for the remainder of this text , we will assume , as before , that the total growth rate @xmath88 is a constant , @xmath1 , and that the presence of the antibiotic modulates selection by increasing the wild type death rate @xmath89 . the logistic growth term makes eq . [ eq : cofxandt ] for the wild type density @xmath15 mathematically equivalent to eq . [ eq : survival - probability0 ] for the establishment probability @xmath17 and hence , all results carry over with minor modifications ( eqs . [ eq : step-1 ] , [ eq : positive - u ] , and [ eq : quasi - static ] ) . in particular , the predicted steady - state population density @xmath15 has a broad tail in the case of a step - like antibiotic profile , and closely follows the mirror image of the antibiotic profile if this is shallow . note , however , that we are unable to observe the broad tail in simulations because of the discreteness effects associated with the small population size at the front ( see si section 7 ) . given an explicit wild - type population profile , it is plausible to assume that the selective pressure felt by the mutants is purely due to competition with the wild type , i.e. , @xmath93 , such that the local establishment probability of the mutants is coupled to the wild type profile . this should indeed be the case if the antibiotic does not directly alter the birth and death rate of the mutants . hence , if @xmath15 changes sharply , i.e. , on length scales shorter than @xmath65 , we expect to see the signatures of a step - like antibiotic profile also in the behavior of @xmath17 . if , on the other hand , @xmath15 changes slowly in space , then the quasi - static approximation may become applicable . the case of a step - like antibiotic concentration profile @xmath94 for illustrates both scenarios ( see fig . [ fig : driftcu]a ) : for @xmath95 , @xmath15 approaches 1 exponentially ( eq . [ eq : positive - u ] ) , such that @xmath17 is well - described by the broad tail in eq . [ eq : step-1 ] . for @xmath96 , instead @xmath15 decays slowly and the quasi - static approximation can be used such that @xmath97 . we compare the simulations with numerical and approximate analytical solutions for @xmath15 and @xmath17 in detail in the si , section 7 . the solutions for @xmath15 and @xmath17 can then be used to compute the rate of adaptation @xmath20 ( eq . [ eq : establishment - rate ] ) and total death rate @xmath19 . asymptotically , we find , @xmath98 and @xmath99 which agrees well with the numerical result in the two limiting cases ( fig . [ fig : driftbrq]a ) . the inverse scaling of the establishment score @xmath20 with the gradient steepness @xmath78 is also in agreement with greulich et al . @xcite in relatively steep gradients where the rate of adaptation ( which is proportional to the establishment score defined here ) is dominated by the time until a mutation arises and establishes . hermsen @xcite also finds that the rate of adaptation increases for shallower ( but still relatively steep ) gradients in what is identified as the `` mutation - limited '' regime , where mutations are rare . however , both hermsen and greulich consider the establishment of many mutation ( potentially simultaneously ) , giving rise to a second , `` dispersion - limited '' , regime @xcite in very shallow gradients , in which the mutational supply is large and the rate of adaptation is dominated by the speed with which established mutations invade previously uninhabitable territory . since we follow the fate of individual mutations , our model operates in the mutation - limited regime exclusively , and we therefore do not expect agreement between our and their models . ( magenta ) and establishment probability @xmath17 ( green ) for a step - like antibiotic concentration ( a ) and a very broad concentration profile ( @xmath22 , c ) , for values of the convection speed @xmath100 ranging from @xmath101 to @xmath26 in steps of @xmath102 ( dark to light colors ) . curves were obtained by numerically integrating eqs . [ eq : drift - eq - c ] and [ eq : drift - eq - u ] ( _ methods _ ) . for high negative convection speeds ( dark colors ) , an analytical approximation is valid ( b , si section 5 ) . ( d ) the cut - off @xmath103 of the population density profile in shallow gradients found numerically ( dots ) agrees well with the theoretical prediction , eq . [ eq : cut - off ] ( solid lines).,scaledwidth=49.0% ] so far , we have neglected external flows and only considered random diffusion of individuals in space . in the following , we explore the influence of convection on the population density , the establishment probability , and finally , the treatment efficiency , which , in the absence of convection , is naturally constrained to a relatively small range , see eqs . [ eq : b - analytical ] and [ eq : r - analytical ] . when a population is subjected to convection , it will generally move in the direction of the flow , leading to either a depletion or enrichment of both mutant and wild types in the antibiotic region ( unless convection is too strong , see si section 4 ) , depending on whether the flow is directed away ( counter - flow ) or towards ( co - flow ) the gradient , respectively . we define the direction of the flow such that a negative convection speed @xmath13 points in the positive @xmath16-direction , towards the gradient ( see fig . [ fig : sketch ] ) . the corresponding equations for @xmath104 and @xmath105 are @xmath106 where @xmath107 . note the change in sign for the convection term in the equations for @xmath108 and @xmath109 that stems from the reverse time direction in @xmath105 . figure [ fig : driftcu ] shows the effect of convection on both the population density @xmath109 and the establishment probability @xmath108 , for a step - like antibiotic profile and a shallow gradient ( @xmath22 ) , obtained by first evaluating @xmath109 from eq . [ eq : drift - eq - c ] and then solving @xmath108 from eq . [ eq : drift - eq - u ] using the steady - state density profile ( see _ methods _ ) . consider first the step - like case . there is a strict distinction between positive and negative convection : negative convection ( darker colors in fig . [ fig : driftcu]a and c ) tends to broaden the profiles away from the step and increase both population density and establishment probability near the step . in fact , it can be shown that for our model both profiles decay asymptotically as @xmath110 far from the step in the limit of strong negative convection ( fig . [ fig : driftcu]b , si section 5 ) . by contrast , positive convection away from the antibiotic ( @xmath111 ) reduces the wild - type population density in the antibiotic region as well as the establishment probability , giving rise to an exponential decay @xmath112 for @xmath113 . the effects of a finite gradient steepness , characterized by the length scale @xmath70 , can be judged by rescaling eq . [ eq : drift - eq - c ] by @xmath70 , as in eq . [ eq : alpha1 ] , i.e. , @xmath114 for @xmath115 in shallow gradients , the profiles are barely affected ( see fig . [ fig : driftcu]c ) . for @xmath111 , on the other hand , convection can alter the profiles strongly : as @xmath116 , the solution to eq . [ eq : alpha2 ] first becomes insensitive to the diffusion term and finally also to the convection term , such that the population density profile @xmath109 follows the gradient almost perfectly until it drops steeply to zero at the cut - off position @xmath103 ( fig . [ fig : driftcu]d , see si section 6 ) , given by @xmath117 this cut - off position @xmath103 captures the behavior of the numerical solution very well ( fig . [ fig : driftcu]d ) . the establishment probability @xmath108 mirrors @xmath109 as long as @xmath118 : for @xmath62 , we have @xmath119 , and for @xmath59 , we have @xmath120 . for @xmath121 , @xmath108 increases very sharply from 0 to 1 at the cut - off position , leaving only a very small overlap with @xmath109 . hence , the establishment score @xmath20 in shallow gradients becomes very small for high positive convection speeds . and the total death rate @xmath19 in our analytical model . panel ( a ) shows @xmath20 and @xmath19 as a function of the gradient length scale @xmath68 . for @xmath122 , the total death rate @xmath19 approaches the theoretical predictions , eq . [ eq : b - analytical ] . negative / positive convection shifts both @xmath19 and @xmath20 up / down , respectively . plotting @xmath19 and @xmath20 as a function of @xmath100 in panel ( b ) , we recover the same trend as in stepping stone simulations : negative / positive convection increase / decrease both @xmath20 and @xmath19 similarly . ( c ) the treatment efficiency @xmath45 for a step - like antibiotic profile ( black ) and gradients of different steepness , from shallow ( purple ) to steep ( red ) . ] given the scaling of the population density and establishment probability profiles , we can rationalize the behavior of the total death rate @xmath19 and the establishment score @xmath20 . as shown in fig . [ fig : driftbrq]a and b , both @xmath19 and @xmath20 increase for negative convection because @xmath109 and @xmath108 maintain broader profiles . in shallow gradients , @xmath19 and @xmath20 exhibit a plateau , since @xmath109 and @xmath108 are affected only weakly by negative convection ( green ) . for positive convection , both @xmath19 and @xmath20 decrease rapidly by up to 2 order of magnitude for our range of parameters . in shallow gradients , the establishment score @xmath20 decreases faster than the total death rate @xmath19 with increasing convection speed because convection affects both @xmath109 and @xmath108 equally , such that their overlap decreases more rapidly than @xmath109 alone ( see si section 6 ) . however , for the steepest gradients , @xmath19 and @xmath20 decrease at the same pace until at high convection speeds , the death rate @xmath19 decreases more quickly than @xmath20 because of the hard cut - off in @xmath123 ( fig . [ fig : driftbrq ] ) . the behavior of the treatment efficiency @xmath45 ( fig [ fig : driftbrq]c ) follows directly from these considerations . for @xmath115 , i.e. , for convection into the antibiotic , @xmath24 changes only slightly with @xmath100 , and only in relatively steep gradients ( @xmath124 ) . this is in contrast to the case of convection away from the antibiotic : for shallow gradients , i.e. , @xmath125 , @xmath24 is approximately equal to 1 for small positive and all negative convection speeds . for strong positive convection , the establishment score @xmath20 becomes very small , such that @xmath24 increases rapidly ( see si fig . s2 and si section 6 ) . as the step - like case is approached , i.e. , for @xmath126 , the treatment efficiency @xmath24 is _ reduced _ further for large positive @xmath100 , where the total death rate @xmath19 decays more quickly than the establishment score @xmath20 . our numerical findings , supplemented by analytical calculations rationalize all major observations from our simulations . in particular , we find that convection can strongly alter both the total death rate @xmath19 and the establishment score @xmath20 , but in a manner that depends on the steepness of the gradient . although there are subtle differences between the simulated and numerical profiles of @xmath127 and @xmath128 ( see si section 7 ) , our simple model reproduces even the complex phenomenology of the treatment efficiency @xmath24 , including the shift from an increase in @xmath24 in shallow gradients to a decrease in @xmath24 in steep ones . in order to fight an infection efficiently , an antibiotic must kill as many wild - type bacteria as possible before resistant mutants arise and survive against genetic drift . if a mutant arises in a low - concentration region , it has little advantage over the wild type and likely goes extinct . by contrast , a mutant arising in a high - concentration background will quickly establish , thus creating a resistant population that can populate regions in the antibiotic gradient inaccessible to the wild type . high population densities in high - concentration regions induce significant wild - type death , as the susceptible wild type can not survive well in these regions . on the other hand , high population densities in high - concentration regions also maximize the number of resistance mutations that occur in a favorable environment . we have studied this trade - off between drug - induced wild - type death and adaptation in an antibiotic gradient using simulations and analytical theory . our simulations reveal that drug - induced death is highest when the wild - type population density is enriched in high - concentration regions , i.e. , for convection towards the antibiotic , and shallow gradients . however , since each wild - type individual harbors the potential for a resistance mutation to occur , which would then have a big advantage over the wild type in the region where it occurs , the rate of establishment of resistance mutations has the same general behavior . similarly , for strong counter - flow , the drug - induced total death rate is decreased because the wild - type population density is low in high - drug regions , which in turn limits the rate of establishment of new resistance mutations . thus , making ad hoc predictions about the treatment efficiency is difficult . our detailed analysis of population density @xmath15 and establishment probability @xmath17 profiles over a wide range of gradient steepness and convection speeds shows that in shallow gradients , where both @xmath15 and @xmath17 are strongly affected by convection away from the antibiotic , treatment can become an order of magnitude more efficient than in no - flow scenarios . our simulations are a strongly simplified model of real populations . for instance , we have assumed that mutation establish independently , i.e. , we have neglected clonal interference . our predictions are thus strictly speaking only valid when mutations arise rarely enough that they do not interact . typically , this condition is quantified by demanding that @xmath129 , where @xmath130 is the effective population size . in a spatial scenario , however , potentially interfering mutations can only arise in a spatial region where both the population density @xmath127 and the establishment probability @xmath128 are large . thus , our results remain valid as long as @xmath131 . for resistance mutations with small target sizes , mutation rates can be very small , typically less than @xmath132 @xcite , and thus our approximation may be accurate even for relatively large local populations . in addition , even when mutation rates are not small , we expect convection and spatial gradients to have the same qualitative effects on the establishment of resistance mutations , and thus our results should remain qualitatively correct in this case . individuals in our simulation merely occupy `` space '' in their specific deme ; in reality , bacteria have a finite size , and a population front can advance through mere growth , even against strong counter - flow @xcite . conversely , for strong co - flow , individuals may de - adhere and be carried away from the bulk population , thus founding extant colonies that enjoy large growth rates in the absence of competition for resources . such processes can be studied by generalizing the diffusion term in our model to a long - range dispersal term as used frequently to model epidemics @xcite . since long - range dispersal can allow individuals far from the population front to escape the bulk population , we expect it to increase the total establishment probability and thus the rate of adaptation relative to short - range dispersal as discussed here . we have only discussed one - dimensional populations here , but real surface - bound microbial populations typically grow as two - dimensional colonies , with complex spatial patterns . the establishment of beneficial mutations in microbial colonies has recently been discussed @xcite . due to the particular strength of genetic drift at the front of such populations , beneficial mutations first have to reach a threshold size ( depending on the strength of the selective advantage ) neutrally before they become established . once the mutant clone reaches the threshold size , the selective advantage of the mutants can deterministically drive them to fixation in the population . during the initial phase , the mutant clone is contained between boundaries with characteristic stochastic properties that are not captured in our one - dimensional model @xcite . however , if the threshold size is small , the boundary fluctuations will not have a large impact on the growth of mutant clones . in such cases , we expect our results to apply also to two - dimensional populations . the emergence of drug resistance remains a topic of significant interest , both from a scientific and a public health point of view . considerable effort is brought forward to create novel antibiotics @xcite and new therapy strategies are developed that attempt to limit the emergence of resistance @xcite , but more research is needed to understand how resistance evolves in complex spatio - temporal settings like the spatial gradients discussed in this paper . in particular , as we have shown here , convection constitutes an important factor in shaping the adaptation to antibiotics in spatial concentration gradients and should receive more attention from both theorists and experimentalists . research reported in this publication was supported by the national institute of general medical sciences of the national institutes of health under award number r01gm115851 , by a national science foundation career award and by a simons investigator award from the simons foundation ( o.h . ) . the content is solely the responsibility of the authors and does not necessarily represent the official views of the national institutes of health . the simulations were run on the savio computational cluster resource provided by the berkeley research computing program . suppose particles branch at rate @xmath133 and disappear at rate @xmath134 . then the probability @xmath135 that a particle or its offspring survive until time @xmath136 satisfies @xmath137u(0,t)+\epsilon a \left[1-\left(1-u(0,t)\right)^2\right]\ ] ] the first term on the right - hand side describes the case of neither disappearing nor branching . the second one accounts for the fact that when the initial particle branches then there are two particles , and the probability of survival of at least one lineage is @xmath138 minus the square of both lineages disappearing . using time - translation invariance @xmath139 , we obtain in the limit @xmath140 @xmath141 to extend this equation for spatial degrees of freedom simply account in eq . [ eq : survival - prob1 ] for the random jumps to neighboring lattice sites in a small time interval @xmath142 . [ [ solving - the - establishment - probability - for - step - like - concentration - profile ] ] solving the establishment probability for step - like concentration profile ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the ( ultimate ) establishment probability of a mutation in step - like antibiotic concentration profile given by @xmath143 satisfies eq . 7 in the main text , repeated here for completeness : @xmath144 to solve this equation , we first treat both sides of the step independently , and then fix integration constants by invoking differentiability at @xmath145 . the fixation probability of mutants in the region @xmath59 ( where @xmath146 is determined by the following differential equation : @xmath147 with boundary conditions , @xmath148 and @xmath149 . we exploit a mechanical analogy to solve this equation . we write eq . [ eq : safe - haven - dynamics ] as @xmath150 with a ` potential energy ' @xmath151 . we can then determine the solution for @xmath108 as @xmath152 where the ` total energy ' @xmath153 was chosen to satisfy the boundary condition as @xmath154 . the solution thus follows as @xmath155 implying @xmath61 for large @xmath72 . in the region @xmath62 the new mutations have a net growth rate @xmath156 . the fixation probability is then given as the solution of @xmath157 exploiting the same mechanical analogy , we have the integral : @xmath158 where we have used the condition @xmath159 to set the total energy . we obtain @xmath160 - 1\right\ } , \label{eq : positive - usi}\ ] ] where @xmath161 . for a shallow gradient where @xmath165 , it follows that @xmath166 and hence @xmath44 . for a step , @xmath19 can be obtained directly by simple integration of @xmath109 , @xmath167 . to compute @xmath20 , the integral for @xmath59 is easily performed to yield @xmath168 , but for @xmath62 , the full solution for @xmath109 can not be easily integrated . however , @xmath169 is well approximated by @xmath170 . thus , we obtain the result @xmath171 , where @xmath172 is the exponential integral function . @xmath24 then evaluates to @xmath173 close to the numerical value of @xmath174 . a population growing logistically with an homogeneous rate across demes leads to a fisher wave with a front velocity given by @xmath175 @xcite . in the presence of a convection @xmath13 , far away from the gradient the wild - type population would generate a front advancing at speed @xmath176 . hence , intuitively , if convection is too strong , i.e. , @xmath177 , the population will not be able to establish a steady - state profile . to show this formally , we first transform eq . 17 by introducing a new variable @xmath178 , such that the convection term is replaced by a death term with effective death rate proportional to @xmath179 , @xmath180 around @xmath145 , the exponential factor is approximately equal to 1 and the nonlinearity @xmath181 can be neglected because @xmath182 . the resulting linear equation then has solutions that can be written as @xmath183 . each mode then satisfies a differential equation @xmath184 with @xmath185 . thus , the lowest mode @xmath186 decays to zero as long as @xmath187 , and all higher modes are suppressed by the coupling of modes through the nonlinearity @xcite . therefore , @xmath188 , and hence @xmath109 , undergo an extinction transition for @xmath189 when @xmath190 . ( eq . [ eq : negconvection - c ] ) and @xmath108 ( eq . [ eq : negconvection - u ] ) matches the simulation results well . ( b ) treatment efficiency @xmath24 for different gradients @xmath78 , with asymptotic result , eq . [ eq : asymptoticq ] ( black ) . ] for negative @xmath13 , we first solve for @xmath109 . for @xmath191 , we can neglect diffusion and find the equilibrium population density by solving @xmath192 the solution for @xmath59 is @xmath193 we thus uncover another characteristic length scale @xmath194 , which quantifies the strength of convection relative to diffusion : if @xmath195 , convection dominates over diffusion , and vice - versa . plugging eq . [ eq : negconvection - c ] into the equation for the establishment probability @xmath108 , eq . 18 in the main text , we find analytical solutions in two cases : for small but finite negative @xmath100 , the convection term and the diffusion term can both be neglected and we recover the quasi - static result @xmath198 . for strong negative convection , @xmath199 , we can neglect only the diffusion term and integrate directly . the result is @xmath200 since @xmath201 , it is easy to see that this approximation does not yield a finite establishment score for infinite domains since mutants arising far away from the antibiotics region still have a relatively high chance of establishing . in practice , the establishment score is always finite because for finite @xmath100 , the decay is slightly faster that @xmath202 . assuming a finite domain ( or a cut - off , e.g. , because of a finite carrying capacity ) ranging from some @xmath203 to @xmath204 , we can estimate both the establishment score @xmath20 and the total death rate @xmath19 , and thus the treatment efficiency @xmath45 , as @xmath205\\ b=&2\nu \log[1 + \xi_0/2\nu]\\ q=&\frac{\log[1 + \xi_0/\nu]}{2 \log[1 + \xi_0/2 \nu]}\gtrsim \frac{1}{2}\end{aligned}\ ] ] where we have assumed @xmath206 in the last step . in our numerical evaluations , @xmath207 and @xmath208 , such that we have @xmath209 . . both in the region of high antibiotic concentration @xmath210 and no antibiotic concentration @xmath211 , the population density approaches its asymptotic value exponentially ( ( a ) and ( b ) , analytical results as dotted lines ) . fitting the numerical curves with the analytical expressions , eqs . [ eq : posxi ] and [ eq : negxi ] , we obtain the apparent exponential prefactors @xmath212 , which match well with the expressions for @xmath213 given in the text . ] positive convection ( @xmath111 ) away from the antibiotic limits diffusion of the wildtype into the antibiotics , thus leading to a reduced wildtype population density in the region of antibiotics . for strong convection and a step - like drug profile , the population density decays exponentially over the characteristic distance @xmath214 . this follows by considering the steady - state equation for @xmath215 , @xmath216 for @xmath59 , @xmath217 is positive while @xmath218 is negative ; because these two terms can not add up to zero , we can not neglect the diffusive term . however , since we expect @xmath109 to be small , we can instead neglect @xmath217 to obtain the equation @xmath219 , such that @xmath220 where @xmath221 , which approximates the numerical solution well ( fig . [ fig : step - posv ] ) . for @xmath62 , the corresponding equation reads @xmath222 which does not have a simple analytical solution . however , we can find by inspection that @xmath109 approximately behaves as @xmath223 where @xmath224 is parametrized well by @xmath225 ( see fig . [ fig : step - posv ] . for @xmath226 and @xmath227 , we can transform eq . 18 by defining @xmath228 to obtain @xmath229 convection only has a relatively small effect on @xmath109 and @xmath108 , and only in the region of strong antibiotics ( where @xmath230 ) , as long as @xmath231 . for large enough @xmath72 this will no longer be satisfied . when @xmath232 , the convection term leads to a cut - off . assuming the cut - off is faster than exponential , we can neglect the non - linearity near the cut - off and thus find a transition to negative growth rate @xmath233 for @xmath234 . solving for @xmath103 , we obtain the approximate cut - off position @xmath235 note that @xmath236 for @xmath237 . for @xmath238 , we can estimate @xmath20 and @xmath19 analytically by approximating @xmath239 and similarly for @xmath108 . this leads to@xmath240 for @xmath241 , this agrees well with the numerical result for shallow gradients ( @xmath242 ) , as shown in fig . [ fig : si_compare ] . when @xmath121 , i.e. , when @xmath243 , @xmath109 and @xmath108 retain a small overlap , as shown in fig . 4b in the main text . hence , the treatment efficiency @xmath24 does not diverge at @xmath244 . ( left column , magenta ) and establishment probability @xmath108 ( right column , green ) , for three different antibiotic gradients ( step - like , ( a ) and ( e ) ; intermediate gradient @xmath23 , ( b ) and ( f ) ; shallow gradient @xmath22 , ( c ) and ( g ) ) . deviations between simulations and our numerical model can be traced back to a difference in effective convection speed : the population density feels a slightly higher effective convection @xmath245 ( d ) , while the establishment probability profile reflects the actual convection speed in the simulation @xmath246 ( h ) until the deviations in @xmath109 become too large at high @xmath246 . ] and establishment score @xmath20 . deviations appear mainly at high positive convection speeds @xmath247 in shallow gradients , where the wave front is characterized by strong fluctuations ( see fig . [ fig : comparisonsimnum ] and si section 7 . ( b ) our theory reproduces the observed trends of the treatment efficiency @xmath24 except for the highest positive convection speeds . ] our model for the dynamics of the population density , like the original fisher equation , ignores the discrete nature of individuals . this has been shown to have important consequences , particularly at the tip of the population front , where the number of individuals per deme is small . a heuristic way to implement the discreteness of individuals consists in introducing a cut - off in the growth rate when the local population size becomes too small @xcite . such a cut - off alters the front profile , and the resulting correction to the expansion velocity decays only slowly with the carrying capacity @xmath0 , @xmath248 in our simulations , we use @xmath249 , which should lead to a correction to the velocity by almost 25% . quantitative comparison between numerics and simulations reveals such correction . in shallow gradients with positive convection , where the population density and establishment probability profiles are effectively cut off at a characteristic position @xmath103 ( see eq . [ eq : cut - offsi ] ) , we compute the effective convection speed by comparing the cut - offs in simulated profiles with the numerical curves . the results are shown in fig . [ fig : comparisonsimnum](d ) and ( h ) : the population density obtained from simulations at convection speed @xmath246 has cut - offs positions corresponding to a numerical convection speed @xmath245 that is shifted up by a constant . for instance , even at @xmath250 , we observe a cut - off consistent with a convection speed @xmath251 . the same shift is not visible in the apparent convection speeds obtained from the establishment profiles ; in this case , @xmath252 , except for very high velocities , where @xmath108 is entirely determined by the population density profile @xmath109 ( which deviates from the analytical prediction , see fig . [ fig : comparisonsimnum]c ) . despite these subtle deviations between our model and simulations , we expect good agreement as long as the population density is not too small ; since the establishment probability @xmath17 is typically low where the wildtype density @xmath15 is high and vice - versa , and the establishment score @xmath20 depends only on the product of @xmath15 and @xmath17 , regions of small population density should not significantly impact the integral in most cases . [ fig : comparisonsimnumbr ] shows that there is indeed good agreement between the establishment score @xmath20 , the total death rate @xmath19 and the treatment efficiency @xmath24 obtained from simulations and from numerical evaluation of our model . ling s , hu z , yang z , yang f , li y , chen q , gong q , wu d , li w , tian x , hao c , hungate e a , daniel v t , hudson r r , li w h , lu x , ling s , hu z , yang z , yang f , li y , lin p , chen k , dong l and cao l 2016 _ proceedings of the national academy of sciences _ * 113 * e663e663
since penicillin was discovered about 90 years ago , we have become used to using drugs to eradicate unwanted pathogenic cells . however , using drugs to kill bacteria , viruses or cancer cells has the serious side effect of selecting for mutant types that survive the drug attack . a key question therefore is how one could kill as many cells as possible for a given acceptable risk of drug resistance evolution . we address this general question in a model of drug resistance evolution in spatial drug gradients , which recent experiments and theories have suggested as key drivers of drug resistance . importantly , our model also includes the influence of convection , resulting for instance from blood flow . using stochastic simulations , we quantify the trade - off between the killing of wild - type cells and the rise of resistance mutations : shallow gradients and convection into the antibiotic region promote wild - type death , at the cost of increasing the establishment probability of resistance mutations . we can explain these observed trends by modeling the adaptation process as a branching random walk . our analysis also reveals that the trade - off between death and adaptation depends on a few key control parameters , which compare the strengths of the spatial drug gradient with the convection and random dispersal terms . our results show in particular that convection can have a momentous effect on the rate of establishment of new mutations , and may heavily impact the efficiency of antibiotic treatment .
discovery of superconductivity led to an enormous growth of research in both fundamental science and technology . while these materials are extensively used in medical , scientific and engineering tools , potential application in lossless power transmission , fast transport system etc . has been the major driving force in the search of new superconducting materials with high transition temperature . in this respect , the discovery of superconductivity in fe - based systems @xcite renewed great attention due to the complex interplay of magnetism and superconductivity.@xcite these materials are significantly different from copper oxide superconductors.@xcite it is believed that fe 3@xmath3 states derive the exotic properties in these systems unlike cuprates , where the ligands play significant role . among fe - based superconductors , fe(tese ) group of compounds are believed to be the most correlated ones due to their large ` chalcogen height'@xcite ( the height of the anions from the fe - plane)@xcite . these materials form in anti - pbo - type crystal structure ( space group @xmath4).@xcite the parent compounds , fete exhibits a spin density wave ( sdw)-type antiferromagnetic transition at 65 k @xcite and fese is a superconductor below 8 k.@xcite homovalent substitution of te at se - sites introduces disorder in the system that is expected to reduce the superconducting transition temperature , @xmath5.@xcite in contrast , @xmath5 increases in this system with the maximum @xmath5 of 15 k for about 60% of te concentration,@xcite where disorder is expected to be the most prominent . excess fe ( fe@xmath6te@xmath7se@xmath8 ) in these materials exhibits magnetic ordering with bicollinear commensurate / incommensurate structure that survives even in the highest @xmath5 compound as a short range order.@xcite electronic structure studies show band narrowing and large mass enhancement due to electron correlation.@xcite the normal phase ( above @xmath5 ) of these compounds is complex exhibiting deviation from metallic conductivity near 300 k similar to that often observed in small gap semiconducting systems . lowering of temperature leads to a transition to metallic conductivity around 150 k and eventually superconductivity appears below 15 k. evidently , fe(tese ) exhibits plethora of complexity such as large electron correlation , curious disorder effects , complex normal phase etc . placing them in the pathway between other fe - superconductors and cuprates . here , we studied the electronic structure of fete@xmath0se@xmath1 employing high resolution photoemission spectroscopy and band structure calculations . our results reveal anomalous temperature evolution of the spectral function when the probing photon energy is varied that could be explained in terms of matrix element effect in photoemission and correlation induced enhanced chalcogen @xmath2-contributions near fermi level . the observed change in electrical resistivity around 150 k has also been found to influence the photoemission spectra . the single crystalline sample of fete@xmath0se@xmath1 @xcite was grown by flux method and characterized by @xmath9-ray diffraction , laue , mussbauer and tunneling electron microscopic measurements establishing stoichiometric and homogeneous composition of the sample with no trace of additional fe in the material . the photoemission measurements were carried out using a r4000 wal electron analyzer from gammadata scienta , monochromatic al @xmath10 ( @xmath11 = 1486.6 ev ) , he i ( @xmath11 = 21.2 ev ) and he ii ( @xmath11 = 40.8 ev ) photon sources and an open cycle helium cryostat , lt-3 m from advanced research systems . the sample was cleaved at a base pressure better than 4@xmath1210@xmath13 torr at each temperature just before the measurements . the energy resolutions were fixed to 2 mev , 5 mev and 350 mev for he i , he ii and al @xmath10 energies , respectively . band structure calculations were carried out using full - potential linearized augmented plane wave method using wien2k software.@xcite in local density approximations ( lda ) , local uniform charge densities are considered to calculate the exchange correlation functionals . to improve upon this approximation , the exchange correlation is treated within the generalized gradient approximation ( gga),@xcite where the gradient terms of the electron density are added in the exchange correlation functionals . the gga+@xmath14 method @xcite ( @xmath14 = on - site electron correlation strength ) was employed for @xmath3 electron interactions . the effective on - site coulomb repulsion strength , @xmath15 = 0 - 7 ev ( @xmath16 ; @xmath17 = hund s exchange integral ) was used to visualize the correlation induced changes in the electronic structure . 2@xmath122@xmath122 supercell was used to generate the structure of fese@xmath18te@xmath18 . a ( 25@xmath1225@xmath127 ) @xmath19-point mesh corresponding to 364 @xmath19 points in the irreducible part of the brillouin zone in the monkhorst - pack @xcite scheme was used during the self - consistent cycle . spectra of fe(tese ) ( open circles ) and fe metal ( closed circles - adapted from ref . lines show the fit to the experimental spectrum . ( b ) fe 2@xmath2 spectra at 300 k and 10 k. ( c ) expanded view of the temperature evolution of fe 2@xmath20 peak . ( d ) valence band spectra at 300 k and 10 k. ( e ) expanded view near fermi level region . ] in fig . 1(a ) , we show the fe 2@xmath2 core level spectra exhibiting interesting evolution with temperature and unusually large asymmetry relative to that of fe metal.@xcite the low energy excitations across the fermi level in the photoemission final state lead to asymmetry in the spectral lineshape of metallic systems . however , the significantly large asymmetry in the present case requires consideration of three asymmetric features as denoted by 1 , 2 and 3 in fig . these features can be attributed to the well screened final state ( peak 1 ) and the other final states corresponding to different electron - hole excitations as observed in earlier studies of other related systems.@xcite while the peak 1 is distinctly visible due to better energy resolution in this study , its intensity relative to the higher binding energy features is weaker than that in fe indicating poorer itineracy of the conduction electrons in fete@xmath0se@xmath1 . interestingly , the intensity of the peak 1 increases gradually with the decrease in temperature ( see figs . 1(b ) and 1(c ) ) . earlier theoretical studies@xcite based on dynamical mean field theory showed that in a strongly correlated system , the valence band spectra exhibit two features . one feature appears at the fermi level called coherent feature , which represents the itinerant electrons . the second feature , termed as incoherent feature or lower hubbard band , corresponds to the correlation induced localized states and appears at higher binding energy . decrease in temperature leads to an enhancement of the coherent feature intensity with a consequent decrease in the incoherent feature intensity . the valence band spectra in figs . 1(d ) and 1(e ) exhibit scenario similar to the above theoretical results ; the intensity near the fermi level gradually increases with the decrease in temperature relative to the intensities at higher binding energies . this behavior of the spectral function is a manifestation of the correlation induced effect and also suggests that the intensity of itinerant electrons increases with the decrease in temperature . the enhancement of the well screened feature in the fe 2@xmath2 core level spectra ( peak 1 ) with the decrease in temperature is consistent with this scenario . = 4.0 ev ) and the experimental spectrum . ( b ) fe 3@xmath3 , ( c ) se 4@xmath2 and ( d ) te 5@xmath2 pdos from gga ( thin line ) and gga+@xmath14 ( thick line ) calculations with @xmath14 = 4.0 ev . ] in order to investigate the correlation induced effects further , the calculated electronic density of states are compared with the experimental spectra in fig . 2 . finite electron correlation leads to a spectral weight transfer from @xmath21 to higher binding energies leading to an enhancement of intensity around 2 ev . since the transferred spectral weight appears in the energy range similar to the energy range of bonding bands , distinct identification of the incoherent feature is non - trivial in this system . we observe that the calculated spectral functions in the whole valence band energy range with finite @xmath14 exhibit better description of the experimental spectrum . for example , the calculated spectrum for @xmath224 ev is shown in the figure providing a good description of the experimental features . it is well established that covalency plays significant role in the electronic structure of various ( 3@xmath3 , 4@xmath3 amd 5@xmath3 ) transition metal compounds@xcite . energy bandwidth depends of the degree of hybridization with the neighboring sites . therefore , electron correlation would affect the electronic states with different orbital character differently depending on their degree of itineracy in the uncorrelated system . this is evident in fig . 2 exhibiting significant transfer of the fe 3@xmath3 partial density of states ( pdos ) to higher binding energies . subsequently , the se 4@xmath2 / te 5@xmath2 contributions increase near @xmath21 . these results suggests strong renormalization of fe @xmath3 - chalcogen @xmath2 covalency in this system . and rescaled @xmath2-states [ ( te 5@xmath2 + se 4@xmath2 ) @xmath12 2 ] for different values of @xmath14 . the contribution of @xmath2-states near the fermi level gradually increases with respect to the fe 3@xmath3 contributions with the increase in @xmath14 . ] the change in @xmath2 and @xmath3 contributions with the increase in @xmath14 is demonstrated in fig . the electronic states near @xmath21 is primarily dominated by the fe 3@xmath3 contributions for @xmath14 = 0.0 ev . with the increase in @xmath14 , the fe 3@xmath3 partial density of states gradually shifts towards higher binding energies with substantial increase in intensity around 2 ev . subsequently , the chalcogen @xmath2 contributions near @xmath21 gradually increases making the covalent mixing much stronger compared to the uncorrelated case . , and ( g ) @xmath23 & fit range . ] the high resolution spectra exhibit curious evolution with temperature . the he ii spectra in fig . 4(a ) exhibit a broad feature near @xmath21 with no distinct change in lineshape with temperature apart from a sharp fall in intensity at @xmath21 due to the superconducting gap below @xmath5.@xcite in contrast , the he i spectra exhibit two distinct features at 0.1 and 0.3 ev . the intensities of these peaks increase gradually with the decrease in temperature . since the spectra were collected using transmission mode with a large acceptance angle of @xmath2415@xmath25 , the experimental spectra will be a representation of the total density of states . thus , the change in spectral features might have a origin different from momentum dependence . the photoemission cross section for fe 3@xmath3 states is 4.833 and 8.751 at 21.2 ev and 40.8 ev photon energies , respectively . while that of as 4@xmath2 states is 3.858 and 0.2949 , respectively . thus , the relative photoemission cross - section@xcite of as 4@xmath2 to that of fe 3@xmath3 states enhances from 0.034 at 40.8 ev to 0.8 at 21.2 ev . since the valence states are well described by the linear combination of atomic orbitals , one can conclude that the contribution of @xmath2 orbital character in the conduction electronic states is significant and exhibit higher degree of sensitivity to the change in temperature . the spectral density of states , sdos at different temperatures are estimated by the division of the high resolution spectra by the energy resolution broadened fermi - dirac function . the representation of the 10 k spectra is a convolution of the above sdos by the resolution broadened fermi function of 10 k. thus obtained 10 k spectra based on 300 k and 50 k spectra are compared with the raw data at 10 k in fig . the 300 k spectra exhibit large intensity at @xmath21 . the intensity gradually decreases with temperature as also seen in the sdos obtained by symmetrization ( @xmath26 ) in fig . such decrease in carrier concentration at @xmath21 with the decrease in temperature independent of the analysis procedure adopted might be the origin for the absence of metallic resistivity behavior near 300 k although the density of states indicate metallic phase . the photoemission spectra , however , exhibit monotonic decrease in intensity at @xmath21 with the decrease in temperature that can not capture the transition to metallic temperature dependence of resistivity below 150 k. there are prediction of the persistence of short range antiferromagnetic order in this composition @xcite that might introduce such anomaly in resistivity . in fig . 4(c ) , the raw data at 50 k and 10 k are observed to be shifted due to the formation of the superconducting gap below @xmath5 . a shift of 5 mev of the 50 k data towards higher binding energy reproduces the rising part of the spectrum at 10 k excellently well suggesting the superconducting gap of about 5 mev . the lineshape of the spectral functions is investigated by fitting the raw data with a polynomial ; the best fit is found for the relation , @xmath27 as demonstrated in fig . the fitting parameters are shown in figs . 4(f ) and 4(g ) . the observed energy exponent of 1.5 indicates importance of magnetic fluctuations in the ground state.@xcite while the fitting range starts limiting itself in the low energy regime below 150 k , coefficient , @xmath23 for the spin fluctuation term starts becoming larger at the same temperature , where the transition to metallic phase occurs indicating their importance in the low temperature properties . effect of disorder has extensively been studied theoretically and experimentally.@xcite altshuler and aronov showed that charge disorder in correlated systems exhibits @xmath28 dependence of the spectral lineshape near fermi level and the spectral intensity at @xmath21 is proportional to the square root of temperature.@xcite in the present case , sdos at @xmath21 , represented by @xmath29 in fig . 4(f ) does not follow the typical square root dependence on temperature in conformity with the above observation indicating importance of magnetic interactions in these systems . in summary , we studied the electronic structure of fe - based superconductor , fete@xmath0se@xmath1 employing high resolution photoemission spectroscopy and high quality single crystalline sample . the core level and valence band spectra exhibit signature of electron correlation induced features . the intensity at the fermi level reduces gradually with the decrease in temperature suggesting decrease in conductivity as manifested by insulating temperature dependence of electronic conduction in transport measurements . the transition from insulating to metallic conductivity could be related to the change in spectral lineshape near the fermi level . electron correlation renormalizes the covalency significantly leading to a significant chalcogen @xmath2 orbital contributions near the fermi level . interestingly , the electronic spectra possessing large @xmath2 character exhibit stronger sensitivity to the change in temperature . the spectral lineshape and its temperature evolution indicate importance of magnetic fluctuations in the electronic properties . evidently , ligands play an important role in high temperature superconductivity as also found in cuprates . the authors n. s. and k. m. acknowledge financial support from the dept . of science and technology , govt . of india under the swarnajayanti fellowship programme . 99 y. kamihara , h. hiramatsu , m. hirano , r. kawamura , h. yanagi , t. kamiya , and h. hosono , j. am . soc . * 128 * , 10012 ( 2006 ) . y. kamihara , t. watanabe , m. hirano , and h. hosono , j. am . soc . * 130 * , 3296 ( 2008 ) . n. kurita , m. kimata , k. kodama , a. harada , m. tomita , h. s. suzuki , t. matsumoto , k. murata , s. uji , and t. terashima , phys . b * 83 * , 214513 ( 2011 ) ; g. adhikary , n. sahadev , d. biswas , r. bindu , n. kumar , a. thamizhavel , s. k. dhar , and k. maiti , j. phys . condens . matter * 25 * , 225701 ( 2013 ) . a. damascelli , z. hussain , and z .- x . shen , rev . mod . phys . * 75 * , 473 ( 2003 ) . k. kuroki , h. usui , s. onari , r. arita , and h. aoki , phys . b * 79 * 224511 ( 2009 ) . y. mizuguchi and y. takano , j. phys . . jpn . * 79 * , 102001 ( 2010 ) . m. tegel , c. lhnert , and d. johrendt , solid state comm . * 150 * 383 ( 2010 ) . f. ma , w. ji , j. hu , z .- y . lu , and t. xiang , phys . rev . lett . * 102 * , 177003 ( 2009 ) . a. subedi , l. zhang , d. j. singh , and m. h. du , phys . b * 78 * 134514 ( 2008 ) . hsu , j .- y . luo , k .- w . yeh , t .- k . chen , t .- w . huang , p. m. wu , y .- c . lee , y .- l . huang , y .- y . yan , and m .- k . wu , pnas * 105 * 14262 ( 2008 ) . song , y .- l . wang , p. cheng , y .- p . jiang , w. li , t. zhang , z. li , k. he , l. wang , j .- f . jia , h .- h . hung , c. wu , x. ma , x. chen , and q .- k . xue , science * 332 * 1410 ( 2011 ) . a. ghosal , m. randeria , and n. trivedi , phys . b * 65 * , 014501 ( 2001 ) ; m. chand , g. saraswat , a. kamlapure , m. mondal , s. kumar , j. jesudasan , v. bagwe , l. benfatto , v. tripathi , and p. raychaudhuri , phys . rev . b * 85 * , 014508 ( 2012 ) . c. s. yadav , p. l. paulose , and k. m. subhedar , euro . phys . lett . * 90 * 27011 ( 2010 ) . w. bao , y. qiu , q. huang , m. a. green , p. zajdel , m. r. fitzsimmons , m. zhernenkov , s. chang , m. fang , b. qian , e. k. vehstedt , j. yang , h. m. pham , l. spinu , and z. q. mao , phys . * 102 * , 247001 ( 2009 ) . k. nakayama , t. sato , p. richard , t. kawahara , y. sekiba , t. qian , g. f. chen , j. l. luo , n. l. wang , h. ding , and t. takahashi , phys . lett . * 105 * , 197001 ( 2010 ) . f. chen , b. zhou , y. zhang , j. wei , h .- w . ou , j .- f . zhao , c. he , q .- ge , m. arita , k. shimada , h. namatame , m. taniguchi , z .- y . lu , j. hu , x .- y . cui , and d. l. feng , phys . rev . b * 81 * , 014526 ( 2010 ) . a. tamai , a. y. ganin , e. rozbicki , j. bacsa , w. meevasana , p. d. c. king , m. caffio , r. schaub , s. margadonna , k. prassides , m. j. rosseinsky , and f. baumberger , phys . lett . * 104 * , 097002 ( 2010 ) . m. aichhorn , s. biermann , t. miyake , a. georges , and m. imada , phys . b * 82 * , 064504 ( 2010 ) . c. s. yadav and p. l. paulose , new j. phys . * 11 * , 103046 ( 2009 ) . p. blaha , k. schwarz , g. k. h. madsen , d. kvasnicka , and j. luitz , * wien2k * , an augmented plane wave + local orbitals program for calculating crystal properties ( karlheinz schwarz , techn . universitt wien , austria ) , 2001 . isbn 3 - 9501031 - 1 - 2 . j. p. perdew , k. burke , and m. ernzerhof , phys . lett . * 77 * , 3865 ( 1996 ) . m. t. czyzyk and g. a. sawatzky , phys . b * 49 * , 14211 ( 1994 ) . j. monkhorst and j. d. pack , phys . b * 13 * , 5188 ( 1976 ) . j. f. van acker , z. m. stadnik , j. c. fuggle , h. j. w. m. hoekstra , k. h. j. buschow , and g. stroink , phys . b * 37 * , 6827 ( 1988 ) . m. takahashi and j .- igarashi , phys . rev b * 85 * , 085128 ( 2012 ) . a. georges , g. kotliar , w. krauth , and m. j. rozenberg , rev . * 68 * , 13 ( 1996 ) ; m. imada , a. fujimori , and y. tokura , rev . mod . phys . * 70 * , 1039 ( 1998 ) . e. pavarini , s. biermann , a. poteryaev , a. i. lichtenstein , a. georges , and o. k. andersen , phys . * 92 * , 176403 ( 2004 ) ; k. maiti , phys . b * 73 * , 235110 ( 2006 ) ; k. maiti , phys . rev . b * 73 * , 115119 ( 2006 ) ; k. maiti , phys . rev . b * 77 * , 212407 ( 2008 ) . v. r. r. medicherla , s. patil , r. s. singh , and k. maiti , appl . . lett . * 90 * , 062507 ( 2007 ) . j. j. yeh and i. lindau , _ at . data and nucl . data tables _ * 32 * , 1 ( 1985 ) ; k. maiti , p. mahadevan , and d. d. sarma , phys . b * 59 * , 12457 ( 1999 ) ; k. maiti and d. d. sarma , phys . rev . b * 58 * , 9746 ( 1998 ) . k. maiti , r. s. singh , v. r. r. medicherla , s. rayaprol , and e. v. sampathkumaran , phys . lett . * 95 * , 016404 ( 2005 ) ; k. maiti , r. s. singh , and v. r. r. medicherla , phys . rev . b * 76 * , 165128 ( 2007 ) ; r. s. singh , v. r. r. medicherla , k. maiti , and e. v. sampathkumaran , phys . b * 77 * , 201102(r ) ( 2008 ) . p. a. lee and t. v. ramakrishnan , rev . phys . * 57 * , 287 ( 1985 ) . b. l. altshuler and a. g. aronov , solid state commun . * 30 * , 115 ( 1979 ) . m. kobayashi , k. tanaka , a. fujimori , s. ray , and d. d. sarma , phys . lett . * 98 * , 246401 ( 2007 ) ; k. maiti , r. s. singh , and v. r. r. medicherla , europhys . letts . , * 78 * , 17002 ( 2007 ) ; _ ibid . b * 76 * , 165128 ( 2007 ) .
we investigate the electronic structure of fete@xmath0se@xmath1 employing high resolution photoemission spectroscopy and _ ab initio _ band structure calculations . fe 2@xmath2 core level and the valence band spectra exhibit signature of strong electron correlation in the electronic structure . the electronic states near the fermi level reduces in intensity with the decrease in temperature in conformity with the insulating transport observed near 300 k. the observation of an insulator to metal transition around 150 k in the transport properties may be related to the spectral lineshape change in the vicinity of the fermi level observed in this study . the spectral features near fermi level exhibit significant @xmath2 orbital character due to the correlation induced fe @xmath3 spectral weight transfer . the experimental spectra reveal dominant temperature dependence of the spectral functions possessing large @xmath2-character . these results demonstrate significant renormalization of the character of the conduction electrons due to electron correlation and emphasizes the importance of ligand states in the superconductivity of these materials .
the observation of high - redshift grbs such as grb970508 , grb971214 , and grb980703 @xcite confirmed the cosmological origin of grbs @xcite . high redshifts ( @xmath23 ) of grbs would imply a much larger apparent energy budget than the conventional energy budget @xmath24ergs unless some significant beaming is involved @xcite . the isotropic energies of observed grbs with known redshifts range from @xmath25ergs to @xmath26ergs . a broad luminosity function can be a reasonable answer for this wide range , and can be naturally expected if we assume beaming of grbs . the suggested correlation between grbs and certain types of supernovae ( ib / c ) implies that the total energy available for grbs could be rather limited at least in the low energy grbs @xmath27erg / sec ( wang and wheeler 1998 ; woosley et al . 1999 , see also kippen et al . 1998 ; graziani et al . 1999 ) . if we consider the supernova explosion as a universal grb source , then the reported grb - supernova correlation implies that the high redshift grbs must be strongly beamed with the beam opening solid angle , which is approximately given by @xmath28 when we consider beaming , a broad luminosity range and a stringent constraint on the intrinsic energy are anticipated . @xcite considered the effect of the luminosity function due to the cylindrical - beam with the inhomogeneous distribution of grb sources on statistics of grbs . he analytically showed that the luminosity function follows a power law with the index @xmath29 in the very narrow beam model when @xmath8 . his result has been confirmed by mao and yi ( 1994 ) and chang and yi ( 2000 ) . @xcite studied the statistical properties of luminosity functions derived using a conic beam with a uniform spatial distribution of grb sources . they examined the cumulative probability distribution of the peak count rates . they found that the maximum redshift increases as a product of the lorentz factor and the opening angle decreases . we attempt to generalize the luminosity function induced by the cylindrical - beam by introducing the conical shape in such a beam . if the progenitors of grbs were produced in the late stage of massive stars @xcite , their spatial distribution may follow the star - formation rate ( sfr ) of massive stars @xcite . we use the recently observed sfr data to derive the possible number density distribution of grb sources @xcite . in this paper , we study how beaming - induced luminosity functions affect the statistics of the observed grbs in the batse 4b catalog @xcite considering the two different spatial distributions of grb sources : an sfr - motivated distribution and the uniform distribution . we employ two different sfr - motivated distributions . we discuss the effect of the photon index @xmath13 ( @xmath30 and 2.0 ) in the conlusion . we assume a flat universe with no cosmological constant , and adopt the hubble constant @xmath31 ( cf . yi 1994 ; mao and yi 1994 ) . we analyze 775 observed bursts in the batse 4b catalog in order to improve the statistical significance . assuming the bimodal distribution of durations of grbs @xcite we divide the total sample into two subgroups according to their duration and study the statistics of each subgroup . the selection criteria applied to observational data are described in section 2 . luminosity functions due to beaming are discussed in section 3 . we present results and discussions in section 4 . we summarise our conclusion in section 5 . we adopt the bursts in the batse 4b catalog @xcite and calculate their @xmath3 for selected grbs @xcite , using fluxes in channels 2 and 3 . we choose bursts detected in 1024 ms trigger time scale with the peak count rates satisfying @xmath32 . applying these criteria , we select 775 bursts among 915 bursts in the batse 4b @xmath33 table . the @xmath3 value for this total sample is 0.3177 @xmath34 0.0102 . then we divide the selected sample into two subgroups according to the burst durations ( @xmath35 ) , which is motivated by the reported bimodal structure in grb durations @xcite . we define bursts with @xmath36 as long bursts , and those with @xmath37 as short bursts . their estimated @xmath3 values are 0.2901 @xmath34 0.0113 ( 588 bursts ) and 0.4178 @xmath34 0.0239 ( 149 bursts ) , for long bursts and short bursts respectively . the number in parentheses indicates the number of bursts belonging to each subgroup . the @xmath3 values obtained from the two subgroups show deviations from the values which are expected for a uniform distribution in a euclidean space . in our subgroups , the long bursts have a larger deviation from the euclidean value . it agrees with the claim by tavani ( 1998 ) . according to @xcite , bursts with long durations ( @xmath38 ) and hard spectra ( @xmath39 ) show the largest deviation from the so - called euclidean value . we assume two kinds of number density distributions of grb sources : ( i ) the sfr - motivated distribution @xmath40 and ( ii ) the uniform distribution @xmath41 . we assume a two - sided gaussian function as @xmath42 @xcite , and fit @xmath42 to the extinction - corrected sfr data @xcite . the observed sfr gradually decreases or even keeps flat over a wide redshift range beyond a certain redshift ( cf . madau et al . the sfr - motivated number density distribution is approximated by @xmath43 , & ( z < z_{\rm c } ) \\ & n_{\rm { \circ , sfr } } ~\exp \bigl [ - { ( z - z_{\rm c})^2 \over{\delta z_2}^2 } \bigr ] , & ( z > z_{\rm c } ) , \nonumber\end{aligned}\ ] ] where @xmath44 and @xmath45 are widths of the fitting functions for the left part and the right part and @xmath46 is a critical redshift . all parameters are determined by fitting the observational data . we set @xmath47 @xmath48 and we also effectively fix @xmath49(z ) to @xmath50 beyond @xmath51 . the proportional constants @xmath50 and @xmath52 are normalized to satisfy the number from the observational data . we consider the beaming - induced luminosity function following @xcite , where the relativistic beam has a constant cross - sectional area , i.e. a perfectly collimated cylindrical beam . under this circumstance , a bulk motion of all the particles emitting the photons is parallel to the cone axis . in this case , the beam opening angle is irrelevant for the luminosity function . we assume that the beamed emission from `` standard bursts '' which have the same intrinsic luminosity @xmath53 . the advantage of this model is that it physically defines a range of the luminosity , i.e. @xmath54 $ ] , for a given photon index @xmath13 , and the lorentz factor @xmath55 . an apparent luminosity function @xmath56 due to the uniformly distributed beam in space is given by @xmath57 where @xmath58 and @xmath59 . we adopt @xmath8 @xcite and 2.0 @xcite and the lorentz factor @xmath60 ( see piran 1999 ) . if we assume a bi - polar jet , the minimum and the maximum luminosities are naturally given as follows : @xmath61 where @xmath62 and @xmath63^{-1}$ ] @xcite . the intrinsic luminosity @xmath53 for each subgroup is determined by the @xmath3 test . the obtained @xmath53 values are @xmath64 for the specific set of adopted beaming parameters . this is a much smaller value than that required in the standard candle model , which is @xmath65 . we calculate two statistical quantities using the beaming - induced luminosity function explained in this subsection . first , we calculate the @xmath3 as a function of threshold flux @xmath66 @xcite . the threshold flux is obtained when we put the maximum detectable redshift @xmath6 in @xmath67 ^ 2}},\ ] ] where @xmath68 is the luminosity , @xmath69 is the speed of light , @xmath70 is the hubble constant , and @xmath13 is the photon index . after we obtain a set of @xmath66 , we calculate the @xmath3 for each threshold flux . the @xmath3 curve converges to 0.5 as @xmath66 increases sufficiently large . we convert flux to the number of bursts and plot @xmath3 as a function of the number of bursts for all bursts and long and short bursts in fig . 1 ( a ) for @xmath8 ( thin lines ) and @xmath10 ( thick lines ) . the @xmath71 vs. number of bursts relation essentially gives the same information that one obtains from the conventional logn @xmath72 logp plot . our @xmath3 test does not include the threshold effect due to the detection efficiency . second , we calculate the fraction of grbs . we define the fraction of bursts located at a redshift larger than @xmath73 as @xmath74 in fig . 1 ( b ) , we plot the fraction of bursts as a function of redshift for each burst subsample for the cylindrical - beam case with different number density distributions . as a more realistic example , we adopt conically - beamed emission from grbs , which was studied by @xcite , and extended by @xcite . we consider the luminosity function produced using three different opening angles , i.e. @xmath75 , and @xmath76 . we calculate @xmath3 and the @xmath15 for each subgroup with the derived luminosity function . the bulk lorentz factor is set to be 100 . according to @xcite , the probability that we observe the bright bursts rapidly increases as the opening angle increases , while relatively dim bursts are not as detectable as brighter ones . therefore , as the opening angle increases , the derived luminosity function becomes similar to that of the standard candle case , in which all bursts have the same maximum luminosity given by this luminosity function . when the opening angle is small @xmath77 , the luminosity function gives the same result as the cylindrical beaming case . using this luminosity function , we calculate the same statistics as we did in previous subsection . results are shown in figs . 2 - 5 for @xmath8 ( thin lines ) and @xmath10 ( thick lines ) in the cases of each sample . the left panels in figs . 1 - 5 ( a ) show the @xmath78 as a function of the number of bursts . we show results obtained from different luminosity distributions and number density distributions . the thin lines in fig . 1 - 4 represent the results obtained with @xmath8 and thick lines @xmath10 . in figs . 1 - 4 , we assume that the sfr - motivated number density distribution , which is flat beyond a redshift larger than the critical redshift , @xmath51 . finally , we show our results obtained with the sfr - motivated number density distribution , which gradually decreases and @xmath8 in fig . 5 ( a ) . in all cases , we assume that n(z ) rapidly increases if @xmath79 . the dotted lines represent the observed @xmath3 curves with the @xmath80 bound . the solid lines indicate the theoretical results derived from our luminosity functions with the sfr - motivated number density distribution of grb sources . we plot the results obtained for the luminosity functions with the uniform distribution of burst sources as dashed lines , for comparaison . apparently all the luminosity functions we have studied satisfy the observed @xmath3 curve at the 3@xmath81 significance level . although the number density distribution of grb sources is believed to correlate with the sfr , we can not rule out the uniform distribution of burst sources in this analysis ( see also krumholz et al . 1998 ) . the right panels in figs . 1 - 5 ( b ) show @xmath82 in terms of @xmath83 . the y - axis corresponds to the fraction of bursts which have redshifts larger than a certain redshift @xmath83 . solid lines result from the sfr - motivated number density distribution for the spatial distribution of grbs and dashed lines the uniform distribution . filled squares in figs . 1 , 3 , and 5 ( b ) locate the known redshifts of 15 observed grbs so far . the redshift values are quoted from http://cossc.gsfc.nasa.gov/batse/counterparts/grb_table.html . we compare the observed redshift values only with the long subsample because all grbs with the known redshifts belong to the long bursts . it is interesting to note that , if we assume @xmath8 , a large fraction of grbs are distributed at high redshifts , provided that grbs are assumed to follow the sfr , regardless of specific beaming models . however , in the case of @xmath10 , @xmath6 is significantly reduced , though it still remains quite large . the fraction for @xmath84 at @xmath85 for a cylindrical - beam case is @xmath86 for long and total samples when @xmath8 . this is too large a value given the fact that the observed grbs with such high redshifts account for @xmath87 of the total bursts . this fraction become @xmath88 when @xmath10 . it is also interesting to note that , though their @xmath3 value is close to a euclidean value , most of the short bursts are distributed at high redshifts in the case of @xmath8 . when we assume that the long bursts are uniformly distributed in space , the @xmath6 estimate for the broad beaming can not explain the largest observed redshift in both cases of @xmath8 and 2.0 . our study is based on an assumption that all grbs have the same intrinsic luminosity and they are all beamed . because the @xmath3 test is not sufficiently sensitive to the observed data , different luminosity functions are essentially indistinguishable . if there is an intrinsic luminosity function and the degree of beaming of grbs is moderate , we may obtain results similar to what we have presented in this study . if all grbs are highly beamed and their sources follow a sfr - like distribution , the maximum detectable redshift becomes very large , even in the case that the observed @xmath3 is sufficiently close to the euclidean value , i.e. , @xmath89 . therefore , it is difficult to rule out a possibility that the apparent euclidean value of @xmath90 may be due to the luminosity function , for instance , induced by beaming . in other words , the so - called euclidean value may have nothing to do with the euclidean distribution . if the total burst sample is a mixture of broad ( or nearly non - beamed ) beams and narrow beams , the exact fraction of the strongly beamed bursts may affect the burst rate estimate significantly . this is beyond the scope of this paper . for a given lorentz factor , the range of the derived beamed luminosity distribution becomes broader as the beam opening angle decreases . we obtain @xmath91 in the cylindrical - beam case . however , in the case of broad beaming , @xmath92 becomes substantially smaller . this is only an effective range because the cut - off for the minimum luminosity given by the conic beam is rather loosely constrained . it is obvious that the luminosity range decreases when the opening angle increases for a fixed lorentz factor . this is simply explained as follows : assuming that we observe a beamed emission with an opening angle @xmath93 , if we define @xmath94 as an angle between the line of sight and the symmetry axis of the conic beam , we can observe the beamed emission only within a certain range of @xmath94 , @xmath95 . for instance , when @xmath96 , @xmath97 is approximately @xmath93 . for a given lorentz factor , @xmath95 increases as the opening angle of cone becomes larger . the apparent luminosity range from a broad beam would decrease gradually with @xmath94 , and in the extreme case , the isotropic radiation would have constant luminosity regardless of @xmath94 . however , when @xmath93 is very small , the apparent luminosity rapidly decreases with @xmath94 , and the derived luminosity range is narrower than a broad - beam case . we have demonstrated that the beaming - induced luminosity function may explain the statistics of the observed grbs . in the case of the cylindrically beamed luminosity function , we determined the intrinsic luminosity @xmath53 by the @xmath3 test for each subsample . in all cases , the intrinsic luminosity is much smaller than that of non - beamed luminosity functions , i.e. @xmath98 . the maximum luminosity obtained from our luminosity functions , in both the cylindrical - beam and the conic - beam cases , is @xmath99 . this is compatible with the required value for isotropic radiation . the ratio between the maximum luminosity for the long bursts and that for the short bursts @xmath100 is @xmath101 for the narrow beam and @xmath11 for the broad beam , respectively . this implies that these two subgroups may have two different intrinsic luminosities , and hence probably two different origin . the obtained @xmath6 in the cylindrical - beam case are @xmath102 ( @xmath8 ) and @xmath103 ( @xmath10 ) for the total selected sample , and long and short subgroup , respectively . the obtained @xmath6 in the broader beam case ( i.e. @xmath104 ) are @xmath105 ( @xmath8 ) and @xmath106 ( @xmath10 ) for the total selected sample , and long and short subgroup , respectively . the maximum known redshift we consider , @xmath107 , is estimated for grb971214 . however , even a larger value of @xmath108 is suggested for grb980329 ( http://cossc.gsfc.nasa.gov/batse/counterparts/[email protected] ) . without any detailed information on the redshift distribution of grbs , the maximum redshift gives a rather strong constraint on the ratio between the lorentz factor and the opening angle @xmath93 . for the conic beam case , the luminosity function derived for the narrow opening angle seems to fit the @xmath3 curve better . from this simple analysis we have carried out , we have shown that the beaming - induced luminosity function may account for the basic statistical properties of the observed grbs . the apparent euclidean distribution of grbs may be an indication of the presence of the grb luminosity function . although it is the beyond the scope of this work , it is interesting to point out that there could be a potentially observable correlation between spectral hardness and luminosity since the most doppler - boosted burst would probe the hardest part of the spectrum . any wide scatters in intrinsic spectral shapes would make the detection of this correlation complicated . we thank h. kim and k. kwak for useful discussions . iy is supported in part by the krf grant no . 1998 - 001-d00365 . we appreciate the referee , ralph a. m. j. wijers , for his suggestions and some useful information .
most of gamma - ray bursts ( grbs ) models have predicted that the intrinsic isotropic energy is limited to below @xmath0ergs . recently claimed high redshifts , correlation with supernovae , and connection to the cosmic star formation activity point to a different energy requirement and source evolution history possibly with strong beaming . we study the effects of the beaming - induced luminosity function on statistics of observed grbs , assuming the cosmological scenario . we select and divide the batse 4b data into 588 long bursts ( t@xmath1 sec ) and 149 short bursts ( t@xmath2 sec ) , and compare the statistics calculated in each subgroup . the @xmath3 of the long bursts is @xmath4 , and that of the short bursts is @xmath5 . for luminosity function models , we consider a cylindrical - beam and a conic - beam . we take into account the spatial distribution of grb sources as well . a broad luminosity function is naturally produced when one introduces beaming of grbs . we calculate the maximum detectable redshift of grbs , @xmath6 . the estimated @xmath6 for the cylindrical - beam case is as high as @xmath7 ( @xmath8 ) and @xmath9 ( @xmath10 ) for the long bursts and @xmath11 ( @xmath8 ) and @xmath12 ( @xmath10 ) for the short bursts where @xmath13 is the photon index . the large @xmath6 value for the short bursts is rather surprising in that the @xmath3 for this subgroup is close to the so - called euclidean value , 0.5 . we calculate the fraction of bursts whose redshifts are larger than a certain redshift @xmath14 , i.e. @xmath15 . when we take @xmath16 and apply the luminosity function derived for the cylindrical - beam , the expected @xmath15 is @xmath17 and @xmath18 for long bursts . when we increase the opening angle of the conic beam to @xmath19 , @xmath15 decreases to @xmath20 at @xmath21 . if we assume @xmath10 , the conic - beam with @xmath19 can not explain the redshift distribution of the observed grbs . we conclude that the beaming - induced luminosity functions are compatible with the redshift distribution of the observed grbs although the apparent `` euclidean '' value of @xmath22 might be explained by the standard model .
there are many corrections that one in principle has to take into account when calculating the casimir force between two bodies ; the corrections may come from finite temperatures , finite extensions of the plates used in the experiments , corrugations on the plates , etc . ( recent reviews of the casimir effect can be found in refs . the correction that we will be concerned with in the present paper is the one coming from finite temperatures . for the most part , we will consider the temperature dependent casimir force between a compact sphere of radius @xmath0 and a plane substrate . the sphere is situated at a fixed distance @xmath1 from the plane ( @xmath1 denotes the minimum distance between the surfaces ) . the sphere and the substrate are assumed nonmagnetic , but we consider the case where they may be made from different materials . we will moreover assume the proximity force theorem @xcite to hold ; this means that @xmath1 must be much less than @xmath0 . ( for corrections to this see refs . @xcite . ) on experimental grounds it is evidently desirable to calculate the casimir forces in a realistic way . we will here take advantage of the excellent numerical dispersive data to which we have access for the materials gold and copper ( and also aluminum ) ( courtesy of astrid lambrecht and serge reynaud ) . we know how the permittivity @xmath2 varies with imaginary frequency @xmath3 over seven decades , @xmath4 $ ] rad / s . we use these data to calculate the forces at two different temperatures , namely at room temperature , @xmath5 k , and at @xmath6 k. the latter temperature is conveniently attainable numerically , and it can for all practical purposes be identified with zero temperature . ( the lowest temperature that we actually tested was @xmath7 k. if t becomes lower , we leave the frequency domain for our numerical dispersion data . it turns out that there are very small deviations between calculated values of the force for @xmath6 k and for @xmath7 k. ) we obtain in this way a realistic picture of the finite temperature correction for these materials . it ought to be emphasized that we are not adopting the so - called modified ideal metal ( mim ) model , which assumes unit reflection coefficients for all but the transverse electric ( te ) zero mode [ see eq . ( [ 10 ] ) below ] ; rather , we are using real data together with the assertion ( based on thermodynamical and electrodynamical arguments ) that the te zero mode is absent , as that is an isolated point which can not be extracted from data alone . our approach is that which we have followed in other recent papers @xcite . the absence of a te zero mode contribution to the casimir effect for a real metal was discussed in detail in ref . @xcite , and also in ref . the result of this assumption , for instance , in the mim model is the presence of a linear temperature term in the expression for the casimir force between two planes , in the limit where @xmath8 . the calculation for a real metal yields a linear temperature correction for low , but not too low temperatures , so that for very low temperatures the force and the free energy have zero slope . by contrast , in the conventional ( old ) model for an ideal metal ( i m ) the te zero mode is included , and it implies that this linear temperature term is omitted . we ought to stress here that at @xmath9 the mentioned difference between a mim and an i m model goes away , as the contributions from the zero frequency te mode as well as from the zero frequency tm mode become buried in an integral over imaginary frequencies from zero to infinity . the experiments of immediate interest for the present theory are the atomic force microscopy ( afm ) tests , performed in particular by mohideen _ et al . _ a point that we ought to emphasize here is that previous analyses have most likely overestimated the accuracy of the afm experiments . thus the recent paper of chen _ et al . _ @xcite , which is based upon a reanalysis of the experiment of harris _ et al . _ ( listed in @xcite ) , claims an over - all experimental precision to be at the 1 percent level . in that apparatus a gold - coated polystyrene sphere mounted on a cantilever of an afm was brought close to a metallic surface and the deflection was measured . however , as discussed in refs . @xcite , at the very short distance of 62 nm ( the minimum distance ) the force at @xmath10 differs from the force at @xmath11 nm by more than 3.5 pn ( the experimental uncertainty claimed by the authors ) when @xmath12 is larger than a few angstroms . this means that @xmath1 should have been measured with atomic precision in order to correspond to the accuracy claimed . as for temperature corrections , these were found in @xcite to be negligible , this being related to their acceptance of the plasma dispersion relation for the material . as the absence of the zero frequency te mode has been controversial , we give in sec . [ sec4 ] a discussion of this in view of recent work by bezerra _ et al . _ @xcite they argue that this mode should be present . in sec . [ sec5 ] we give additional support to our arguments by showing that for a pair of anisotropic polarizable particles the casimir force can vanish in certain directions as the temperature increases towards @xmath13 , and although there are regions of negative entropy connected with the casimir effect , there is no indication that thermodynamics is violated . violation of thermodynamics is used as an argument by bezerra _ et al . _ @xcite to require the presence of the te zero mode . the inadequacy of using a surface impedance approach without including transverse momentum dependence is briefly reviewed in sec . [ si ] . in this paper we put @xmath14 . let the sphere of radius @xmath0 be nonmagnetic , and have a permittivity @xmath15 . as mentioned , the sphere is situated a distance @xmath1 above a plane substrate ; we let the nonmagnetic substrate have permittivity @xmath16 . according to the proximity force theorem @xcite the attractive force @xmath17 between sphere and plane at temperature @xmath18 can in the limit @xmath19 be given approximatively as the circumference of the sphere times the surface free energy density @xmath20 in the parallel - plate configuration : @xmath21 . following essentially the notation of ref . @xcite we can then write the force as @xmath22 . \label{1}\ ] ] here @xmath23 is the dimensionless quantity @xmath24 , and @xmath25 @xmath26 being the component of @xmath27 parallel to the plates in the parallel - plate configuration . further , @xmath28 with @xmath29 are the matsubara frequencies , and @xmath30 is the dimensionless temperature . superscripts tm and te in eq . ( [ 1 ] ) refer to the transverse magnetic and electric modes . the prime on the sum means that the zero mode has to be counted with half weight . with the conventional lifshitz variables defined as @xmath31 we define the two kinds of @xmath32 s , the reflection coefficients for a single interface , as @xmath33 for each medium 1 and 2 , respectively . if the two media are equal , @xmath34 for each kind of mode , then @xmath35 where @xmath36 are the tm , te coefficients defined in ref . note that @xmath37 , with @xmath38 being the same quantity in the two cases . the permittivities @xmath39 are functions of the imaginary matsubara frequencies @xmath28 . in the general case where the media are dispersive , @xmath40 and @xmath41 depend both on @xmath42 and on the matsubara integer @xmath43 . if the media are nondispersive , @xmath40 and @xmath41 are functions of @xmath42 only , independent of @xmath43 . as a general warning , we mention that the proximity theorem assumed here requires @xmath44 to be very small . thus , within the framework of the optical path method recently considered by jaffe and scardicchio @xcite , believed to be more robust than the proximity approximation , disagreement with the latter approximation was found already when @xmath44 became larger than a few percent , whereas the method they propose agrees accurately with the recent exact numerical result of gies _ et al . _ @xcite . as mentioned in the introduction , we will use accurate numerical data for the variation of @xmath45 with frequency for two different substances : gold and copper . these data refer to room temperature measurements . for gold , the data are shown graphically in refs . @xcite . for frequencies up to about @xmath46 rad / s ( note that 1 ev=@xmath47 rad / s ) , the data are nicely reproduced by the drude dispersion relation @xmath48 where for gold the plasma frequency is @xmath49 ev and the relaxation frequency @xmath50 35 mev . for @xmath51 rad / s the drude curve however lies below the experimental curve . all the dispersive data , of which we are aware , refer to room temperature . now , as we will be interested in the casimir force also at low temperatures , we are faced with the problem of how to estimate the permittivity @xmath52 under such circumstances . numerical trials indicate rather generally that the casimir force is very robust under variations in the input values for the permittivity , but at least this issue is a matter of principle . one possible way to proceed is to write the permittivity as @xmath53 } , \label{7}\ ] ] and make use of the bloch - grneisen formula for the temperature dependence of the electrical resistivity @xmath54 . this problem was discussed in appendix d of ref . one may in this way estimate the temperature relaxation frequency to be , in ev , @xmath55 where @xmath56 k for gold . this formula implies that @xmath57 is somewhat higher for low @xmath18 than for @xmath5 k ( cf . fig . 1 in @xcite ) , when the frequencies are less than about @xmath58 rad / s . it is instructive to compare with the parallel - plate configuration , where it is known that the most important frequencies for the casimir force are lying in the region @xmath59 . for @xmath60 m it corresponds to @xmath61 rad / s . for smaller gap widths where the metallic properties of the medium fade away and its plasma properties become more dominant it follows that the most important frequencies become higher . taking all things together , we expect that the influence from the temperature variation in @xmath62 is rather small . some numerical trials that we have done support this expectation . another important point to be mentioned here is that the bloch - grneisen argument sketched above neglects the effect from impurities . these give rise to a nonzero resistivity at zero temperature @xcite . this fact strengthens our argument for setting the contribution from the casimir force from the te zero mode for a metal equal to zero . the issue has been considered in detail also by sernelius and bostrm @xcite . as one can see from fig . 2 in @xcite , there exists a temperature dependent contribution to the casimir force from the te modes . this temperature dependence occurs for low frequencies at low temperature and extends to higher frequencies with increasing temperature . however , the temperature influence fades away before the the first nonzero matsubara frequency is reached . it is therefore permissible to neglect the temperature dependence in @xmath62 . what remains important in @xmath63 is the constant term @xmath64 that is due to elastic scattering . the consequence is that @xmath65\to 0,\quad\mbox{as}\quad \zeta\to0,\ ] ] so from ( [ 4 ] ) @xmath41 vanishes at @xmath66 . if one neglects this crucial constant @xmath67 , one can end up with a violation of the nernst heat theorem @xcite . see also bostrm and sernelius @xcite for related discussion of these points . therefore , in the following we will restrict ourselves to using the room - temperature values for @xmath63 throughout , even when calculating @xmath68 at different temperatures . even with this simplification , we point out that the temperature dependence in the casimir force turns up in a rather complex way . namely , the temperature occurs at three different places : \(i ) in the prefactor in eq . ( [ 1 ] ) ; \(ii ) in the lower limit of the integral ; \(iii ) in the dependence of @xmath69 and @xmath70 on @xmath18 via the matsubara frequencies in the permittivity : @xmath71 . between a gold sphere and a gold plate versus gap @xmath1 , when @xmath5 k and @xmath72 @xmath73m.,height=377 ] figure [ fig1 ] shows , for a gold sphere and a gold plate , how the attractive force @xmath74 varies with @xmath1 in the interval from about 150 nm to 1 @xmath73 m , when @xmath72 @xmath73 m and @xmath5 k. as mentioned , the empirical data for @xmath2 are directly usable as input in eq . ( [ 1 ] ) . when @xmath75 nm , the force is calculated to be 67.22 pn . a similar calculation can be made for a very low temperature , in order to show the magnitude of the temperature influence in the force . we assume throughout , as mentioned earlier , that @xmath76 mev . using matlab we found @xmath6 k to be a numerically stable and reasonable lower limit . this temperature is moreover low enough to be identifiable with @xmath9 for all practical purposes . we ought to stress that we choose to perform the @xmath77 calculation _ numerically _ , inserting realistic data for @xmath2 . this is in principle different from the conventional @xmath9 calculation for an idealized metal , where one simply puts @xmath78 for all frequencies . ( recall that at @xmath9 the difference between a mim and an i m model goes away , because of the very close spacing between the matsubara frequencies . ) between a gold sphere and a gold plate , versus gap @xmath1 for @xmath72 @xmath73m.,height=377 ] rather than showing the calculated result for @xmath79 explicitly , we show in fig . [ fig2 ] the difference between the forces , @xmath80 , defined as @xmath81 an important property seen from this curve is that @xmath82 is positive . the force is thus _ weaker _ at room temperature than at @xmath9 . this is the same effect as was found in fig . 5 in @xcite . this behavior is thus a consequence of lifshitz formula plus realistic input data for the permittivity ; there are no further assumptions involved . when @xmath75 nm , we find @xmath83= 2.54 pn , which means that the force is reduced by 3.6 percent compared to the @xmath9 case . for larger distances , @xmath84 nm , the temperature effect becomes larger . thus for @xmath84 nm the force is 9.38 pn at @xmath5 k and 10.19 pn at @xmath6 k , yielding a 7.9 percent reduction at room temperature . at @xmath85 m , the corresponding numbers are 0.59 pn at @xmath5 k and 0.73 pn at @xmath6 k , which means a 19 percent reduction . this agrees in magnitude well with the temperature corrections in the case of parallel - plate geometry , as is seen from fig . 5 in ref . @xcite . admitting an error of @xmath86 in the @xmath43 summation in eq . ( [ 1 ] ) , we found the necessary number of terms to be in excess of 34000 in the case of the lowest separation investigated numerically , @xmath87 nm ( not shown in the figure ) . when @xmath75 nm , about 11000 terms were required . at larger separations the necessary number of terms became considerably reduced ; thus the case @xmath60 m corresponded to about 2700 terms . the calculation of the force between a gold sphere and a copper plate gave very similar results . thus for @xmath75 nm the force was 67.19 pn at @xmath5 k and 69.75 pn at @xmath6 k , corresponding to a reduction of 3.7 percent at room temperature . at @xmath88 m , the forces turned out to be the same ( to the accuracy of two decimals ) as in the au - au case . between a gold sphere and a copper plate versus @xmath1 , when @xmath5 k and @xmath72 @xmath73 m . bottom curve is the @xmath66 contribution . , height=377 ] in fig . [ fig3 ] we show , by the full line , how the force between a gold sphere and a copper plate varies with @xmath1 , at @xmath5 k. the curve is calculated from eq . ( [ 1 ] ) , using the empirical data for these two materials directly . our reason for giving this curve anew , in spite of its identity with the curve in fig . [ fig1 ] for all practical purposes , is that we have supplied the following new element , namely at the bottom of the figure we show explicitly the contribution from the @xmath66 term to the force . namely , for the mim model in which @xmath89 for @xmath90 @xcite , we have for the free energy @xmath91 showing the @xmath66 contribution in the first term . thus @xmath92 meaning that the corresponding force contribution @xmath93 becomes @xmath94 from the figure this term contributes an increasing part of the total force for increasing @xmath1 , and gives the full contribution in the classical limit @xmath95 . we would like to make a direct comparison of our prediction with the recent experiment of decca _ et al . _ this experiment measured the casimir force between a gold sphere and a copper plate by means of a microelectromechanical torsional oscillator , for separations in the range 0.22 @xmath73 m . the radius of the sphere was @xmath96 @xmath73 m , thus the same radius as we have assumed above . their measured values are shown , for instance , in fig . 3 in ref . however , the scale of their figure for the total force prevents comparison with our numerical results with sufficient accuracy to draw firm conclusions about the magnitude of the temperature dependence . but fig . [ fig3 ] in ref . @xcite also gives a comparison with theoretical values by plotting the difference . these theoretical values are evaluated at @xmath9 , but as we do , they also use a drude model to obtain the dielectric constant in the limit of low frequencies . then they find a small temperature correction with the same sign as we have theoretically predicted . from fig . [ fig3 ] in ref . @xcite we estimate @xmath97 to be around 1 pn for @xmath75 nm , and to be too small to be discernible for large gaps , @xmath98 nm ( the scatter of the experimental points is considerable ) . this can be compared with our fig . [ fig2 ] , where we calculated @xmath97 to be 2.56 pn for @xmath75 nm and 0.14 pn for @xmath60 m . this reasonably good agreement between experimental and theoretical results is encouraging , and it indicates that our finite temperature calculations are on the right track . we should also mention that ref . @xcite refers to dynamical measurements that are claimed to rule out our results . however , we believe that the systematic theoretical and experimental uncertainties in this measurement are larger than estimated by those authors . in particular , we re - emphasize the uncertainty of measuring the casimir force due to uncertainties in estimating a systematic shift of position as earlier discussed by iannuzzi _ et al . _ @xcite and milton @xcite , as mentioned in the introduction . bezerra _ et al . _ @xcite claim that the drude dielectric function for metals can not be used in the theory for the casimir force . their opinion is that it violates the nernst theorem in thermodynamics and is furthermore ruled out by a recent experiment . instead the plasma relation should be used for the dielectric function . the latter implies the presence of a transverse electric mode at zero frequency besides the static dipole - dipole interaction ; the latter being the zero - frequency limit of the transverse magnetic mode . including such an transverse electric mode adds a term linear in temperature to the casimir force by which it is increased by a factor of two in the high temperature limit . these authors point to the standard theory , sketched above in sec . [ sec2.1 ] where the relaxation parameter @xmath63 varies with temperature @xmath18 and vanishes at @xmath9 . this is then a situation where in principle a transverse electric zero mode might be present in the casimir force although it should not be present according to maxwell s equations of electromagnetism . however , its presence for @xmath99 , but not otherwise , would make physical phenomena discontinuous as then @xmath99 will give something different from taking the limit @xmath100 . also we question whether a possible vanishing of the relaxation parameter at @xmath9 can dictate the behavior for @xmath101 where @xmath102 . ( as a remark we here note that strictly speaking the statistical mechanical derivation is exact only for dielectric functions independent of temperature , i.e. , when induced dipole moments are harmonic oscillators . a @xmath18-dependence reflects anharmonicity . but we do not expect that this is of crucial significance here . ) as a result , bezerra _ et al . _ @xcite conclude that the drude dielectric function is thermodynamically inconsistent and can not be used to calculate the thermal casimir force for metals . however , we on the contrary have shown that it is thermodynamically consistent leading to zero entropy at @xmath9 in accordance with the nernst theorem @xcite . thus , for @xmath101 a negative entropy contribution due to electromagnetic interaction between media is allowed as long as the total entropy is positive . we have earlier studied a simplified model in this regard @xcite . the model consists of three harmonic oscillators . two of them are the analogues of the two media that interact via the electromagnetic field represented by the third oscillator . in the beginning of their sec . iii , bezerra _ et al . _ @xcite state that our derivations were based upon a constant relaxation parameter @xmath63 . however , we did not require such a limitation , only that @xmath103 is finite . and as the authors further note , the value of @xmath63 is commonly very small , but nevertheless finite , at @xmath9 due to impurities , and as a result the entropy becomes zero at @xmath9 as required . then they write that the negative entropy we found at @xmath104k is a violation of the nernst theorem . but such a negative perturbing entropy is not a violation of the nernst theorem since @xmath18 is finite as already discussed above . the further discussion about relaxation time , impurities , and surface impedance does not invalidate the use of a finite @xmath103 . these authors further remark that we consider nonzero wave vectors @xmath105 for which the reflection coefficient @xmath106 . but this does not mean that reflection properties are different for the fluctuation field compared to real photons as we deal with only one such quantity ; and as discussed above , thermodynamics is not violated . ( see sec . [ si ] below for further discussion of @xmath107 dependence . ) they further note that a material with dielectric constant @xmath108 , like a metal , gives a casimir force that decreases with temperature in some interval and thus implies a negative entropy contribution . they conclude that such a material can not exist as it would violate thermodynamics . then they claim that real media with such large @xmath45 are commonly polar for which @xmath45 rapidly reduces to its optical value connected to its electronic polarizability . so with plate ( or plate - sphere ) separation of @xmath109 the casimir force again becomes monotonically increasing with @xmath18 . however , this does not preserve the monotonic character of the casimir force in general because one can just increase the plate separation . non - monotonic behavior will then reappear when this separation exceeds a wavelength corresponding to the relaxation frequency ( where @xmath45 decreases rapidly ) . in sec . iv of their paper , they again conclude that the use of the drude dielectric function in the lifshitz formula violates the third law of thermodynamics ( the nernst heat theorem ) . this conclusion , stated as a rigorous proof , is made on the basis that the relaxation parameter will be much less than the matsubara frequencies . but this is not a rigorous proof as the relaxation parameter will violate this inequality sufficiently close to @xmath9 ( assuming @xmath110 ) . furthermore the drude dielectric function does not predict a linear temperature correction to the casimir force all the way down to zero temperature . as we have shown earlier the casimir force flattens out and becomes independent of temperature at @xmath9 in accordance with thermodynamics @xcite . ( but the sharpness of this flattening increases with decreasing @xmath63 . ) we further remark that the use of the drude dielectric function is consistent with the @xmath111 limit ( for large separation of the plates ) . use of the plasma model ( @xmath99 ) , however , will give a discontinuous jump of the force in this limit . such a jump is not expected for a continuous change in a physical parameter . also for real metals the @xmath45 effectively will be finite due to the finite size of the plane - sphere configuration . ( here one can note that a medium consisting of separate metal spheres of finite size will be like a polarizable medium with finite polarizability and dielectric constant . ) finally the authors in ref . @xcite conclude that use of the drude dielectric function ( [ 6 ] ) is in contradiction with experiment . however , so far as we can see , experiments performed at a single temperature are at present too uncertain to draw conclusions about temperature variations . this seems even more evident from ref . there detailed analysis of experimental uncertainties are performed and various corrections for very short separations are estimated . they find an uncertainty of 1.75% at 95% confidence level for 62 nm separation . this uncertainty increases to 37.3% for 200 nm . earlier it has been remarked by others @xcite that such experiments are very sensitive to accurate determinations of plate separation since the plate - sphere casimir force is proportional to the inverse cube of separation distance . however , the authors assert that they avoid this problem by making a least squares fit of the resulting data by which zero separation is pinpointed with an uncertainty of 0.15 nm . but in view of the uncertainties of the experiment this does not seem to resolve the disputed temperature dependence as heavy weight is put on the shortest separation of 62 nm where the force is by far the largest and changes most rapidly . thus high apparent precision can be obtained for this separation ( 1.75% ) while for larger separations the uncertainties are rapidly increasing until they become larger than the magnitude of the disputed thermal effect . according to the authors of ref . @xcite the thermal effect in dispute is about 1 - 2% for 62 nm . very recently , a paper has appeared @xcite giving the microscopic theory of the casimir effect . these authors convincingly demonstrate that the te zero mode can not contribute , although the plasma model gives such a contribution . as mentioned above the appearance of a negative casimir entropy for metals in a region of nonzero temperatures has been disputed with the claim that it violates the nernst theorem of thermodynamics @xcite . however , as we argue this negative entropy region does not imply violation of thermodynamics since the casimir free energy is a perturbing one and is not the total free energy of two interacting systems . many such examples are known in statistical mechanics , including that of three interacting oscillators discussed in ref . @xcite . to illustrate this point , we here will consider a pair of particles with strong anisotropic polarizability and thus polarizable only in the @xmath112-direction , e.g. , they may be metal needles . the result for a pair of particles with isotropic polarizability is well known and was rederived in a novel way by brevik and hye @xcite . the latter derivation is easily generalized to anisotropic particles . using the path integral formalism the dipole - dipole interaction of the radiation field is given by eq . ( here and below the numeral i refers the the equations of the reference mentioned . ) equation ( i5.9 ) gives the interaction energy @xmath113a_{1\zeta_n}a_{2\zeta_n}^ * \label{h1}\ ] ] with @xmath114 @xmath115 where the matsubara frequency @xmath116 . the dipole radiating fields @xmath117 are given in ( i5.10 ) . hats denote unit vectors . the @xmath118 are unit vectors of the fourier - transformed fluctuating dipole moments along the polymer " path and @xmath119 is the discretized step length along the polymers " that represent quantized particles in the path integral formalism . the casimir free energy times @xmath120 is now obtained from the average of expression ( [ h1 ] ) squared in accordance with ( i3.5 ) . for the isotropic case one has the average @xmath121 where @xmath122 is the polarizability . this follows from eq . ( i5.3 ) . for the strongly anisotropic situation to be considered below one likewise will find ( @xmath112 denotes the @xmath112-component ) @xmath123 as the only nonzero average as polarizations in @xmath124- and @xmath23-directions are zero in the present case . furthermore let the positions of the particles relative to each other be such that the @xmath112-component of @xmath125 equals @xmath126 . with this the @xmath127 will vanish as polarizations are present only in the @xmath112-direction . ( with this relative position the corresponding electric field is transverse to the @xmath112-direction and thus to each of the dipole moments , i.e. , there is no interaction connected to the @xmath127-term . ) thus only the @xmath128-term remains . so with polarizations restricted to the @xmath112-direction eq . ( i5.14 ) turns into @xmath129 or with ( i5.10 ) inserted the free energy is @xmath130 where @xmath131 as given by ( i5.12 ) . as usual eq . ( [ h6 ] ) gives a negative free energy . however , in the classical limit @xmath132 the casimir free energy ( [ h6 ] ) , and thus the corresponding force , both vanish since only the @xmath133 term will contribute . furthermore with this free energy the contribution to the entropy must be negative ( or at least mainly negative ) as @xmath134 , because @xmath20 must have a generally positive slope , but @xmath135 at @xmath9 as it should . most recently , mostepanenko _ et al . _ @xcite , apparently conceding that their arguments favoring the plasma model over the drude model for the dispersion relation for real metals could not be supported either thermodynamically , electrodynamically , or experimentally , have asserted that for real metals one should use in the reflection coefficients in the lifshitz formula not the bulk dielectric permittivity but the surface impedance . indeed there is much to be said for using the latter . however , in may be shown in general that there is in fact no difference between the reflection coefficients computed using either description @xcite . there is a one - to - one correspondence between the permittivity @xmath45 and the surface impedance @xmath136 , which is given by the ratio of the transverse electric and magnetic fields at the surface . this correspondence , however , necessitates in general that both quantities possess dependence on the transverse momentum @xmath137 . as optical data strongly suggest that this dependence is usually negligible in the permittivity , a strong dependence on @xmath137 is required in @xmath136 @xcite @xmath138 the vanishing of @xmath139 at @xmath140 demonstrates again that the te zero mode does not contribute to the casimir force . ( note that this vanishing happens in the drude , but not the plasma model . ) in contrast , ref . @xcite completely disregard this transverse momentum dependence and moreover make an _ ad hoc _ extrapolation from the infrared region to zero frequency @xcite . the inadequacy of neglecting the transverse momentum dependence has been stressed by esquivel and svetovoy @xcite . in this paper we have sharpened our arguments in favor of using real data for the dielectric functions to apply the lifshitz formula to calculate the force between metal surfaces , in particular between a spherical lens and a flat plate . in principle , one can also use the surface impedance to calculate this force , and although optical data is lacking for the latter , such use should yield the same result . in contrast , the procedures advocated in refs . @xcite contain _ ad hoc _ elements and assumptions . we show both by direct computation , and through analogous models , that there is no conflict with thermodynamical principles , in particular with the nernst heat theorem ( the third law of thermodynamics ) . especially important is the demonstration that the entropy necessarily vanishes at zero temperature . claims that experimental limits on casimir forces preclude our temperature dependence @xcite are , in our opinion , not justified , since the accuracy of the current experiments does not match their precision , especially due to the impossibility of determining the shortest separation distances accurately . undoubtedly , it will take dedicated experiments involving different temperatures to reveal the true temperature dependence of the casimir effect . one idea might be to measure the difference between the casimir forces for the same value of @xmath1 at two different temperatures , for instance 300 k and 350 k. such a difference is directly measurable , in principle . to our knowledge this idea was originally proposed by chen _ et al . _ @xcite , and it was further discussed in ref . @xcite . k. a. milton , j. phys . a * 37 * , r209 ( 2004 ) . k. a. milton , _ the casimir effect : physical manifestations of zero - point energy _ ( world scientific , singapore , 2001 ) . m. bordag , u. mohideen , and v. m. mostepanenko , phys . reports * 353 * , 1 ( 2001 ) . j. blocki , j. randrup , w. j. swialecki , and c. f. tsang , ann . ( n.y . ) * 105 * , 427 ( 1977 ) . r. l. jaffe and a. scardicchio , phys . lett . * 92 * , 070402 ( 2004 ) ; a. scardicchio and r. l. jaffe , nucl . b * 704 * , 552 ( 2005 ) . h. gies , k. langfeld , and l. moyaerts , jhep * 06 * , 18 ( 2003 ) . t. emig , europhys . lett . * 62 * , 466 ( 2003 ) ; r. bscher and t. emig , e - print cond - mat/0401451 . j. s. hye , i. brevik , j. b. aarseth , and k. a. milton , phys . e * 67 * , 056116 ( 2003 ) . i. brevik , j. aarseth , j. s. hye , and k. a. milton , in _ quantum field theory under the influence of external conditions _ , edited by k. a. milton ( rinton press , princeton , 2004 ) , p. 54 , e - print quant - ph/0311094 . bo e. sernelius and m. bostrm , in _ quantum field theory under the influence of external conditions _ , edited by k. a. milton ( rinton press , princeton , 2004 ) , p. 82 . u. mohideen and a. roy , phys . lett . * 81 * , 4549 ( 1998 ) ; a. roy , c .- y . lin , and u. mohideen , phys . d * 60 * , 111101(r ) ( 1999 ) ; b. w. harris , f. chen , and u. mohideen , phys . rev . a * 62 * , 052109 ( 2000 ) . f. chen , g. l. klimchitskaya , u. mohideen , and v. m. mostepanenko , phys . a * 69 * , 022117 ( 2004 ) . d. iannuzzi , i. gelfand , m. lisanti , and f. capasso , in _ quantum field theory under the influence of external conditions _ , edited by k. a. milton ( rinton press , princeton , 2004 ) , p. 11 , e - print quant - ph/0312043 . v. b. bezerra , g. l. klimchitskaya , v. m. mostepanenko , c. romero , phys . a , * 69 * , 022119 ( 2004 ) . a. lambrecht and s. reynaud , eur . j. d * 8 * , 309 ( 2000 ) . m. khoshenevisan , w. p. pratt , jr . , p. a. schroeder , and s. d. steenwyk , phys . rev . b * 19 * , 3873 ( 1979 ) . m. bostrm and bo e. sernelius , phys . lett . * 84 * , 4757 ( 2000 ) ; bo e. sernelius , phys . * 87 * , 139102 ( 2001 ) ; bo e. sernelius and m. bostrm , phys . lett . * 87 * , 259101 ( 2001 ) . v. b. bezerra , g. l. klimchitskaya , and v. m. mostepanenko , phys . a * 65 * , 052113 ( 2002 ) ; _ ibid . _ * 66 * , 062112 ( 2002 ) . m. bostrm and b. e. sernelius , physica a * 339 * , 53 ( 2004 ) . r. s. decca , d. l. lpez , e. fischbach , and d. e. krause , phys . rev . lett . * 91 * , 050402 ( 2003 ) . r. s. decca , e. fischbach , g. l. klimchitskaya , d. e. krause , d. lpez , and v. m. mostepanenko , phys . d * 68 * , 116003 ( 2003 ) . i. brevik and j. s. hye , physica a * 153 * , 420 ( 1988 ) . b. geyer , g. l. klimchitskaya , and v. m. mostepanenko , phys . a * 67 * , 062102 ( 2003 ) . r. esquivel and v. b. svetovoy , phys . a * 69 * , 062102 ( 2004 ) . see also v. b. svetovoy , cond - mat/0412123 . f. chen , g. l. klimchitskaya , u. mohideen , and v.m . mostepanenko , phys . * 90 * , 160404 ( 2003 )
the temperature dependence of the casimir force between a real metallic plate and a metallic sphere is analyzed on the basis of optical data concerning the dispersion relation of metals such as gold and copper . realistic permittivities imply , together with basic thermodynamic considerations , that the transverse electric zero mode does not contribute . this results in observable differences with the conventional prediction , which does not take this physical requirement into account . the results are shown to be consistent with the third law of thermodynamics , as well as being not inconsistent with current experiments . however , the predicted temperature dependence should be detectable in future experiments . the inadequacies of approaches based on _ ad hoc _ assumptions , such as the plasma dispersion relation and the use of surface impedance without transverse momentum dependence , are discussed .
one of the major problems of quantum information is the evaluation of the capacity of quantum communication channels , i.e. , the evaluation of the amount of _ classical information _ which can be reliably transmitted by _ quantum states_. this problem has been extensively studied along with the development of other aspects of quantum information science . early works in this direction were devoted , mainly , to memoryless channels for which consecutive signal transmissions through the channel are not correlated . the capacities of some of such channels were determined @xcite-@xcite and it was proven that in most cases their capacities are additive . for gaussian memoryless channels under gaussian inputs , the multiplicativity of output purities was proven in @xcite and the additivity of the energy - constrained capacity even in the presence of classical noise and thermal noise was proven in @xcite . in the last few years , much attention was given to quantum channels with memory @xcite-@xcite with the hope to find a possibility to enhance the channel capacity by using entangled input states . this would be possible if the capacity of such channels is superadditive . for some of such channels , for example , for bosonic memory channels in the absence of input energy constraints , the additivity conjecture was proven leaving no hope to enhance channel capacity using entangled inputs @xcite . however , for several other examples of quantum channels , where memory is introduced by a correlated noise , it was shown that entangling two consecutive uses of the channel enhances the overall channel capacity . these examples include qubit pauli channels @xcite and bosonic continuous - variable gaussian channels @xcite-@xcite . for qubit channels it was shown that if the noise correlations are stronger than some critical value , maximally entangled input states enhance the channel capacity compared to product input states . for gaussian channels also , entangled states perform better than product states . however , for each value of the noise correlation parameter , there exists an optimal degree of entanglement ( and not maximal entanglement ) that maximizes the channel capacity . quantum channels with correlated noise in dimensions @xmath1 were not considered in the literature in this context yet ( see note added ) . they correspond to a kind of intermediate systems between the qubit and gaussian channels . therefore , we expect to find new features that this intermediate dimensionality can add to the known facts . we start with a short review of the results on the capacities of the pauli qubit channels with memory studied in papers @xcite . then , we propose a generalization of these channels to @xmath0 dimensions and present new results on their capacity . the action of a transmission channel on an initial state @xmath2 is given by a completely positive ( cp ) map @xmath3 @xmath4 the amount of classical information which can be reliably transmitted through a quantum channel is given by the holevo - schumacher - westmoreland bound @xcite as the maximum of mutual information @xmath5 taken over all possible ensembles @xmath6 of input states @xmath7 with _ a priori _ probabilities @xmath8 . the mutual information of an ensemble @xmath6 is defined as @xmath9\ ] ] where @xmath10 $ ] is the von neumann entropy . if we find a state @xmath11 which minimizes the output entropy @xmath12 and replace the first term in ( [ i ] ) by the largest possible entropy given by the entropy of the maximally mixed state , we obtain the following bound @xmath13 where @xmath2 is a @xmath14 matrix . this bound ( [ bound ] ) is very useful for further evaluation because it was shown to become tight for two - qubit channels @xcite which are covariant under pauli rotations , @xmath15 , such that the following equality holds : @xmath16 the two - qubit channels represent two consecutive applications ( denoted by @xmath17 ) of a one - qubit channel . the proof of the above statement is simple . it is based on the fact that the pauli matrices form an irreducible representation of a heisenberg group . this implies that for any two - qubit state @xmath2 , an equal probability ensemble @xmath18 is maximally mixed . now , let @xmath11 be an input two - qubit state that minimizes the output entropy @xmath19 . then , the equal probability ensemble ( [ ens ] ) made from this @xmath11 has the following properties . on the one hand , it is maximally mixed and therefore , it maximizes the first term in ( [ i ] ) . on the other hand , due to the channel covariance ( [ cov ] ) , the second term in ( [ i ] ) becomes equal to the entropy @xmath20 which is minimal by our choice of @xmath11 . therefore , the equal probability ensemble ( [ ens ] ) made from @xmath11 attains the bound ( [ bound ] ) . finally , in order to find @xmath21 for covariant two - qubit channels ( [ cov ] ) it is enough to find one of the _ optimal _ states @xmath11 and calculate the right hand side of the bound ( [ bound ] ) , which is tight in this case . we will present a @xmath0-dimensional version of these arguments in the next section . in general , being a cp map , any quantum channel can be represented by an operator - sum : @xmath22 when @xmath23 represents @xmath24 uses of the same quantum channel , the hilbert space of the initial states is a tensor product such that @xmath25 . the amount of information transmitted per use of the channel in the limit of infinitely many uses of the channel determines the channel capacity @xmath26 if consecutive uses of the channel are not correlated then the repeated uses ( @xmath24 times ) of this memoryless channel can be represented by a tensor product of the form @xmath27 however , in the presence of correlations between consecutive uses of the channel the representation given by eq . ( [ product ] ) is not valid . as an example , for _ pauli _ channels , whose action is represented by pauli matrices , @xmath24 uses of the channel is in general given by @xmath28 where the probability distribution @xmath29 is normalized @xmath30 and its particular properties determine different kinds of @xmath24-qubit pauli channels . a pauli channel is memoryless and can be represented as in eq . ( [ product ] ) when @xmath29 is factorized into probabilities which are independent for each use of the channel , @xmath31 however , if each use of the channel depends on the preceding one , such that @xmath29 is given by a product of conditional probabilities @xmath32 a memory effect is introduced . indeed , the correlations between the consecutive uses of the channel act as if the channel `` remembers '' the previous signal and acts on the next one using this `` knowledge '' . this type of channels is called a markov channel as the probability ( [ pn ] ) corresponds to a markov chain of order 2 . in the two - qubit pauli channels considered in @xcite @xmath33 a memory effect is introduced by choosing probabilities @xmath34 which include a correlated noise represented by the term with a kronecker s delta : @xmath35 the memory parameter @xmath36 $ ] characterizes the correlation `` strength '' . indeed , for @xmath37 , the probabilities of two subsequent uses of the channel are independent , whereas for @xmath38 , the correlations are the strongest ones . these channels are not exactly markov for the following reason . let us send @xmath39 qubits through such channels one after another and consider @xmath24 consecutive pairs of these qubits . the actions of the channel on the qubits belonging to the same pair are correlated according to ( [ mtqc ] ) and ( [ p ] ) . however , the actions of the channel on the qubits from different pairs are uncorrelated so that @xmath40 . therefore @xmath39 consecutive uses of the channel are factorized into pairs and can be represented by a product of two - qubit channels @xmath41 where @xmath17 is defined by ( [ mtqc ] ) . however , even these `` limited '' pairwise correlations lead to the advantages of using entangled input states as it was shown in @xcite . 0.5 cm _ a. quantum depolarizing ( qd ) channel_. this channel is determined by equal probabilities for all pauli matrices except for the identity : @xmath42 it is characterized by the parameter @xmath43 $ ] , which is called `` shrinking factor '' @xcite . 0.5 cm _ b. quasi - classical depolarizing ( qcd ) channel_. this channel is determined by a probability parameter which has the same value @xmath44 for the two pauli matrices which do not introduce any bit - flip and another value @xmath45 for the other two pauli matrices including a bit flip : @xmath46 we call this channel quasi - classical because , it is equivalent to the concatenation of a fully - dephasing channel ( where the quantum phase is lost ) followed by a classical channel akin to the depolarizing channel , where the bit is left unchanged with some probability or is shifted with the complimentary probability . for this channel there is no `` shrinking factor '' but it is nevertheless characterized by a single parameter @xmath47 $ ] . the results of the evaluation @xcite show that product states of the form @xmath48 maximize the mutual information of memoryless channels ( @xmath49 ) of both types , whereas maximally entangled states of the form @xmath50 maximise the mutual information for the strongest noise correlations ( @xmath38 ) . moreover for each channel there exists a crossover point @xmath51 below which ( @xmath52 ) the mutual information is maximized by product states and above which ( @xmath53 ) maximally entangled states maximise the mutual information . this crossover point @xmath51 depends on the probability parameters @xmath44 in ( [ p1 ] ) and ( [ p2 ] ) . in fig . 1 , we plot the mutual information of equal probability ensembles of the type ( [ ens ] ) for product states and for maximally entangled states in the qcd channel ( @xmath54 ) . following the above arguments we evaluate the mutual information according to the right hand side of eq . ( [ bound ] ) using a product state or a maximally entangled state for @xmath11 this plot is qualitatively the same as the one presented in @xcite for the quantum depolarizing channel . as a function of the memory parameter @xmath55 for qcd channel ( @xmath56 ) with @xmath54 . the solid line corresponds to maximally entangled input states while the dashed line corresponds to product states.,scaledwidth=45.0% ] pauli qubit channels are generalized to @xmath0-dimensional heisenberg channels @xcite constructed with the help of `` error '' or `` displacement '' operators acting on @xmath0-dimensional states . @xmath57 here the index @xmath58 characterises a cyclic shift in the computational basis by analogy with the bit - flip , and the index @xmath24 characterises a phase shift . the displacement operators form a heisenberg group @xcite with commutation relation @xmath59 two uses of a @xmath0-dimensional channel are described by @xmath60 by using the notation @xmath61 we emphasize again that ( [ emn ] ) represents two consecutive uses of a channel each acting on @xmath0-dimensional states . for simplicity , we shall call these channels @xmath0-dimensional . we introduce a correlated noise by a markov - type probability @xmath62 here @xmath55 is the memory parameter like in the 2-dimensional case , but the kronecker s delta @xmath63 in ( [ p ] ) is represented now by a product of two kronecker deltas representing the noise correlations separately for displacements ( index @xmath58 ) and for phase shifts ( index @xmath24 ) . in addition , we introduce both phase correlations ( @xmath64 ) as well as phase anticorrelations ( @xmath65 ) with a new parameter @xmath66 characterising the type of phase correlations in the channel . for @xmath56 such a distinction disappears as phase correlations @xmath64 and phase anticorrelations @xmath65 coincide , but , interestingly for higher dimensions the type of phase correlations affects the channel capacity . in the limiting case of infinite dimensional bosonic gaussian channels , only phase anticorrelations were shown to provide some enhancement of the channel capacity due to entanglement @xcite . the examples of two - qubit channels presented above are generalized to @xmath0-dimensions as follows : 0.5 cm _ a. quantum depolarizing ( qd ) channel_. this channel is given by the following probability parameters which are equal to each other for all possible actions of the channel except for the identity as in the 2-dimensional case . @xmath67 the sum of all probabilities is normalised therefore , @xmath44 and @xmath45 are not independent and the channel may be characterized by a single parameter @xmath68 with the range @xmath69 $ ] . this reminds us the``shrinking factor '' for two - qubit qd channel . 0.5 cm _ b. quasi - classical depolarizing ( qcd ) channel_. this channel is given by the following probability parameters @xmath70 the probabilities of the shifts by @xmath58 are equal regardless of the phase shift determined by @xmath24 . moreover , we have chosen them to be equal if @xmath71 . the probability of `` zero '' displacement ( @xmath72 ) differs form others as in the 2-dimensional case . the sum of all probabilities is normalised so that @xmath44 and @xmath45 are not independent and the channel may be characterized by one parameter , which we choose to be @xmath73 because it takes the values in the interval @xmath74 $ ] , which is close to the the``shrinking factor '' for qd channel . for these @xmath0-dimensional channels , we prove , first , that bound ( [ bound ] ) is tight . let @xmath11 be the input state that minimizes the output entropy @xmath75 and let us notice that owing to the commutation relation ( [ cr ] ) the channel determined by ( [ emn ] ) and ( [ pmn ] ) is covariant with respect to the displacements ( [ u ] ) : @xmath76 using the fact that the von neumann entropy is invariant under unitary transformations @xmath77 we come to @xmath78 due to the fact that the group of matrices @xmath79 is an irreducible representation of the heisenberg group , the ensemble of input states @xmath80 taken with equal probabilities provides a maximally mixed output state , which up to the factor @xmath81 is : @xmath82 then using equal probability @xmath81 we insert ( [ mm ] ) into the first term of ( [ i ] ) and ( [ se ] ) into the second term of ( [ i ] ) and thus conclude that this ensemble indeed maximises the mutual information ( [ i ] ) and provides the capacity @xmath83 thus we have proven that in order to determine the capacity of these channels we have to find an _ optimal _ state @xmath11 that minimizes the output entropy @xmath75 . by analogy with the two - dimensional case @xcite we start looking for the optimal @xmath11 by using as an ansatz for the input state a pure state @xmath84 where @xmath85 this allows us to go from a product state to a maximally entangled state by changing the parameters @xmath86 and @xmath87 . indeed , the choice @xmath88 and @xmath89 results in a product state whereas the choice @xmath90 and @xmath89 results in a maximally entangled state . taking into account the form ( [ u ] ) of the displacement operators @xmath91 , the probability distribution @xmath92 ( [ pmn ] ) and the probability parameters @xmath93 for both channels ( [ qdc ] ) and ( [ qcdc ] ) , we evaluate the action of the channel given by eq . ( [ emn ] ) on the initial state @xmath94 in the form ( [ iniphi ] ) . then , we diagonalize the output states and find their von neumann entropy , which allows us to obtain the mutual information according to eq . ( [ irho ] ) . first , we evaluate the action of the qd channel given by eqs . ( [ emn]-[qdc ] ) on a pure initial state given by eq . ( [ iniphi ] ) . the result is given by the following equation @xmath95\ ] ] where factors @xmath96 , @xmath97 , @xmath98 , and @xmath99 for _ even _ dimensions @xmath0 are given by @xmath100 & + & dq(p - q)\sum_{j=0}\alpha_j^2 \nonumber \left(i\otimes|j\rangle \langle j| + |j\rangle \langle j|\otimes i\right ) , \nonumber \\ b & = & d q \sum_{j , m=0}^{d-1}\alpha_j^2 |j+m \rangle |j+m\rangle \langle j+m| \langle j+m| \nonumber \\ & + & d \sum_{j , m=0}^{d-1}\alpha_j\alpha_{j+\frac{d}{2 } } e^{i\left(\phi_j-\phi_{j+\frac{d}{2}}\right)}\nonumber \\ & \times & |j+m \rangle |j+m\rangle \left\langle j+m+\frac{d}{2}\right|\left\langle j+m+\frac{d}{2}\right|,\nonumber \\ c & = & d q \sum_{i , j , m=0}^{d-1 } \alpha_i \alpha_j e^{i(\phi_i+\phi_j ) } \nonumber \\ & \times & |i+m \rangle |i+m\rangle\langle j+m|\langle j+m|,\nonumber \\ [ 2ex ] d & = & ( p - q)|\psi_0\rangle\langle\psi_0 |.\end{aligned}\ ] ] for the initial product state ( @xmath88 ) the output state given by eq . ( [ eqdc ] ) is diagonal and the eigenvalues are found easily as @xmath101 observing that these eigenvalues do not depend on @xmath66 we conclude that product states do not feel the difference between the types of phase correlations in the channel . for the maximally entangled initial state ( @xmath90 and @xmath89 ) we had to rearrange the @xmath102 matrix representing the output state eq . ( [ eqdc ] ) in order to find the eigenvalues @xmath103 \nonumber\\ & + & \mu(2(1-\nu)dq+\nu d^2q+p - q ) , \nonumber \\ \lambda_{m0 } & = & ( 1-\mu)q(1+p - q)+2\mu(1-\nu)dq,\nonumber \\ & & 0 < m < \frac{d}{2 } \nonumber \\ \lambda_{m0 } & = & ( 1-\mu)q(1+p - q ) , \quad \frac{d}{2 } \le m < d , \nonumber \\ \lambda_{mn } & = & ( 1-\mu)q(1+p - q ) , \quad m , n > 0.\end{aligned}\ ] ] we note that for @xmath56 the dependence on @xmath66 also disappears and we recover the eigenvalues obtained in @xcite . observing that the state indexes in @xmath97 ( [ qdcev ] ) contain the term @xmath104 , which can not exist for @xmath105 dimensions , we expect different results for @xmath105 @xmath0 indeed , although the action of the channel on @xmath2 for _ odd _ dimensions is given by the same eqs . ( [ eqdc ] ) and ( [ qdcev ] ) , @xmath97 in this case is different and given by @xmath106 nevertheless , the eigenvalues for the initial product state ( @xmath88 ) are given by the same eq . ( [ eigenprod11 ] ) . however , for the maximally entangled initial state ( @xmath90 ) , rearranging the @xmath102 matrix representing the output state given by eq . ( [ eqdc ] ) we find the eigenvalues which differ from the case of @xmath107 dimensions . @xmath108 \nonumber\\ & + & \mu((1-\nu)(dq+p - q)+\nu ) \nonumber \\ \lambda_{mn } & = & ( 1-\mu)q(1+p - q)+\mu ( 1-\nu ) d q\delta_{m , n } , \nonumber \\ & & m , n > 0.\end{aligned}\ ] ] the details of the analytical evaluation for quasi - classical depolarizing channel are similar . they are presented in the appendix . in section v we will visualise and discuss these analytic results , but before , in the next section we will give some arguments whether these results provide the _ optimal _ @xmath11 . as function of the memory parameter @xmath55 for qcd channel with @xmath54 for different values of the optimisation parameter @xmath109 . , scaledwidth=47.0% ] the task of finding an _ optimal _ @xmath11 becomes easier for the quasi - classical depolarizing channel because we can restrict our search from the whole space to a certain subclass . in order to show this we note , following @xcite , that the phase averaging operation @xmath110 @xmath111 does not affect the qcd - channel in the sense that @xmath112 hence , if @xmath11 is an _ state then @xmath113 is also an optimal state . therefore we can restrict our search from the whole space @xmath114 to @xmath115 . finally , using ( [ u ] ) , it is straightforward to show that any state from @xmath115 is a convex combination of pure states @xmath116 where @xmath117 restricting our search to the states of the form ( [ state ] ) we reduce the number of real optimization parameters from @xmath118 to @xmath119 , which can still be a large number . in order to reduce this number to 1 , we consider the following ansatz @xmath120 interpolating between the product state ( @xmath121 ) and the maximally entangled state ( @xmath122 ) . using the one - parameter family of input states @xmath123 , in fig . 2 we present the mutual information @xmath124 for different values of @xmath109 . the mutual information is monotonously modified when @xmath109 goes from a product state to a maximally entangled state , whereas the crossover point @xmath51 stays intact . however , we can not guarantee that no other configuration of the parameters @xmath125 and @xmath126 minimises the entropy @xmath127 and provides therefore the maximum of the mutual information . in this section we will analyse the analytic results obtained in section iii for product states and maximally entangled states as candidates for the _ optimal _ @xmath11 . we first consider the qd channel in _ even _ dimensions . in fig . 3 we display the mutual information @xmath128 as a function of the memory parameter @xmath55 for product states and maximally entangled states . these curves are seen to cross . we denote the abscissae of the crossing points as @xmath51 . for small noise correlations , @xmath129 , product input states provide higher mutual information and for higher noise correlations @xmath130 maximally entangled states do , as is observed in @xmath56 . \(a ) ( b ) as function of the memory parameter @xmath55 for different dimensions @xmath131 for the qd entanglement - friendly ( a ) channel ( @xmath132 ) and entanglement - non - friendly ( b ) channel ( @xmath133 ) both characterized by @xmath134 . the solid lines correspond to maximally entangled input states and the dashed lines correspond to the product input states.,title="fig:",scaledwidth=47.0% ] 0.7 cm as function of the memory parameter @xmath55 for different dimensions @xmath131 for the qd entanglement - friendly ( a ) channel ( @xmath132 ) and entanglement - non - friendly ( b ) channel ( @xmath133 ) both characterized by @xmath134 . the solid lines correspond to maximally entangled input states and the dashed lines correspond to the product input states.,title="fig:",scaledwidth=47.0% ] we now consider how the crossover point @xmath51 changes with _ even _ dimension @xmath0 for two types of phase correlations in qd channels . let us notice , first , that due to the kronecker s deltas in ( [ pmn ] ) the correlated part of the channel is given by @xmath135 for @xmath133 and @xmath136 for @xmath132 . from the definition of @xmath91 ( [ u ] ) one can see that @xmath24 determines the phase shift . the correlations corresponding to the same phase shift @xmath24 in both entangled signals is what we call `` phase correlations '' ( it happens when @xmath133 ) and the correlations corresponding to opposite phase shifts @xmath24 and @xmath137 is called `` phase anticorrelations '' ( it happens when @xmath132 ) . for @xmath138 we have intermediate situations . in fig . 3 ( a ) , which corresponds to phase anti - correlations ( @xmath132 ) , we see that with an increasing dimension the crossover points move toward smaller @xmath55 thus widening the interval @xmath139\mu_c,1]$ ] where maximally entangled states provide higher values of the mutual information than product states do . for this reason we call this version of the channel `` entanglement - friendly '' . an opposite effect can be seen in fig . 3 ( b ) , which corresponds to phase correlations ( @xmath133 ) . here the crossover point moves toward higher values of @xmath55 with an increasing dimension of the space of states , thus shrinking the interval @xmath139\mu_c,1]$ ] where maximally entangled states perform better . this version of the channel is thus called `` entanglement - non - friendly '' . the difference between the two types of phase correlations and therefore , between the `` entanglement - friendly '' and `` entanglement - non - friendly '' channels may be seen as a result of the fact that maximally entangled states @xmath140 are the eigenstates of phase anti - correlated products of operators @xmath141 . note that for @xmath56 the difference between the two types of phase correlations disappears , which was discussed in section ii and we recover the result obtained in @xcite . in order to see the effect of the phase correlations for higher dimensions and observe intermediate situations we draw in fig . 4 ( a ) the coordinate of the crossover point , @xmath51 , as a function of @xmath0 for different values of @xmath66 . we see that only strongly anticorrelated phases ( @xmath142 ) provide `` entanglement - friendly '' channels so that with increasing dimension the interval of @xmath55 s that are favorable for entangled states increases . in addition , even for @xmath132 this increase continues only up to certain @xmath0 , after which the interval of @xmath55 begins to shrink with increasing @xmath0 . we also note that @xmath51 depends on the shrinking factor @xmath143 . if we drew the curves shown in fig . 4 ( a ) for higher values of @xmath143 , the slope of the curves would become steeper and the upper ones , corresponding to the `` entanglement - non - friendly '' channel ( @xmath144 ) would cross the level @xmath145 at some @xmath146 . in this case , for higher dimensions , the interval @xmath147 would shrink to zero so that there would be no values of @xmath55 for which entangled input states may have any advantage at all . for the qcd channel the results are similar and therefore we do not present the graphs corresponding to fig . 3 and fig . 4 ( a ) . similarly to the picture drawn in fig . 3 , the anticorrelated phases make the qcd channel also `` entanglement - friendly '' and the correlated phases do the opposite . the dependence of @xmath51 on dimension @xmath0 corresponding to fig . 4 ( a ) is also qualitatively similar . however , for the qcd channel the upper curves corresponding to the `` entanglement - non - friendly '' version of the channel ( @xmath144 ) cross the level @xmath145 at some values of @xmath146 whereas for the qd channel @xmath148 for any @xmath66 for all even dimensions @xmath146 as shown in fig . 4 ( a ) . as a result , we conclude that for _ even _ dimensions , the advantages of entangled states are more essential for low ( but not always lowest ) dimensions , anticorrelated phases , smaller values of @xmath143 and qd channels . \(a ) ( b ) vs. _ even _ ( a ) and _ odd _ ( b ) space dimension for the qd channel with @xmath134 for different values of the parameter @xmath66 characterizing phase correlations . , title="fig:",scaledwidth=47.0% ] 0.7 cm vs. _ even _ ( a ) and _ odd _ ( b ) space dimension for the qd channel with @xmath134 for different values of the parameter @xmath66 characterizing phase correlations . , title="fig:",scaledwidth=47.0% ] the analytic formulas for _ even _ dimensions ( [ qdcev ] ) , ( [ qcdcev ] ) and _ odd _ dimensions ( [ qdcodd ] ) , ( [ qcdcodd ] ) are different . this difference comes from the second term of the expression for @xmath97 in eqs . ( [ qdcev ] ) , ( [ qcdcev ] ) for _ even _ dimensions which is absent in eq . ( [ qdcodd ] ) , ( [ qcdcodd ] ) for _ odd _ dimensions . due to the factor @xmath149 in front of @xmath97 in eqs . ( [ eqdc ] ) and ( [ eqcdc ] ) this difference disappears for @xmath132 , which corresponds to the `` entanglement - friendly '' channels . indeed , for the _ odd _ dimensions , in the case of entanglement - friendly ( @xmath132 ) versions of both qd and qcd channels , the difference from fig . 3 ( a ) is only in the position of the curves whereas qualitatively the curves are similar and we do not show them here . however for `` entanglement - non - friendly '' channels ( @xmath132 ) , the difference does exist and can be seen from comparison of fig . 3 ( b ) ( for _ even _ dimensions ) and fig . 5 ( for _ odd _ dimensions ) . indeed , in fig . 5 the crossover points of the entanglement - non - friendly version ( @xmath133 ) of the _ odd_-dimensional qd channel lay on the vertical line @xmath38 therefore , effectively there is no crossover as @xmath55 can not be larger than 1 . in this case for all @xmath55 the maximally entangled input states do not provide higher values of the mutual information than the product states , which is not the case for _ even _ dimensions . in order to see the effect of intermediate phase correlations , we draw in fig . 4 ( b ) the abscissae @xmath51 of the crossover points , as a function of _ odd _ @xmath0 for the qd channel at various values of the `` friendness '' parameter @xmath66 . the upper horizontal line @xmath150 corresponds to the `` entanglement - non - friendly '' version of the channel ( @xmath133 ) presented also in fig . 5 . for the qcd _ odd_-dimensional channel the picture is similar however , the upper horizontal line @xmath150 is achieved even for a nonvanishing value of the `` friendness '' parameter , @xmath151 . we have considered two types of @xmath0-dimensional quantum channels with a memory effect modeled by a correlated noise . we have shown that for @xmath0-dimensional channels with correlated noise there exist crossover points separating the intervals of the memory parameter @xmath55 where ensembles of maximally entangled input states or product input states provide higher values of the mutual information . this result is the same as in the 2-dimensional case . however it always holds only for channels ( which we call `` entanglement - friendly '' channels ) with a particular kind of phase correlations , namely anticorrelations . for these channels the crossover point move , with increasing @xmath0 , towards lower values of the memory parameter thus widening the range of correlations where maximally entangled input states give better results . for usual phase correlations the situation is opposite , namely , for higher dimensions of the space the crossover point is shifted towards @xmath145 so that only for higher degrees of correlations maximally entangled input states have advantages . in addition , for these `` entanglement - non - friendly '' channels the crossover point completely disappears for higher dimensions so that product input states always provide higher values of the mutual information than maximally entangled input states . therefore we conclude that the type of phase correlations strongly affects this entanglement assisted enhancement of the channel capacity . as function of the memory parameter @xmath55 for different dimensions @xmath152 for qd entanglement - non - friendly channel ( @xmath133 ) with @xmath134.,scaledwidth=45.0% ] in addition , we have observed that the parity of the dimension of the space of initial states makes an important difference in the `` entanglement - non - friendly '' channels . not only the curves of the mutual information vs. the memory parameter for odd dimensions are shifted with respect to the curves for even dimensions , but also the channels become completely `` entanglement - non - friendly '' in all odd dimensions so that maximally entangled input states are always worse than product states . strikingly , the channels with anticorrelated noise do not feel the parity of the space at all . however , any non - vanishing degree of the `` entanglement - non - friendly '' correlations reveals the parity effect . we note that the anticorrelated phases remind the bosonic gaussian channels considered in @xcite where the @xmath44 quadratures are correlated while the @xmath45 quadratures are anticorrelated . however , the existence of the crossover point is a significant difference with the case of the gaussian channels for which each value of the noise correlation parameter determines an optimal degree of entanglement ( different from maximal entanglement ) maximising the mutual information . a challenging problem is to find a link between these results for @xmath0-dimensional channels and the results obtained in @xcite for gaussian channels with finite energy input signals . finally we have presented a parametrisation illustrating a `` monotonous '' deformation of the curves of mutual information vs. the memory parameter , from product states to maximally entangled states . the crossover points stay intact during these deformations which may lead to a threshold type transition maximising mutual information from product states to maximally entangled states in case no other states perform better than the maximally entangled ones . we have to make this stipulation here because a full proof of the optimality of maximally entangled input states is still missing . _ note added : _ recently , we became aware of the work by v. karimipour and l. memarzadeh on channels with memory for @xmath153 @xcite . we acknowledge the support of the european union projects secocq , chic and qap . we evaluate the action of the channel given by ( [ emn ] ) on a pure initial state given by eq . ( [ iniphi ] ) taking into account eqs . ( [ u ] ) , ( [ pmn ] ) , and ( [ qcdc ] ) the result is given by the following equation @xmath154 where factors @xmath96 , @xmath97 , and @xmath98 for _ even _ dimensions are given by @xmath155 & + & ( p - q)|\psi_0\rangle\langle\psi_0|.\end{aligned}\ ] ] here , as in the case of the quantum depolarizing channel , these eigenvalues do not depend on @xmath66 . therefore , the product input states do not feel the difference between the two types of phase correlations in the channel . for the maximally entangled initial state ( @xmath90 ) rearranging the @xmath102 matrix representing the output state given by eq . ( [ eqcdc ] ) , we find the eigenvalues @xmath157 for @xmath56 , the dependence on @xmath66 disappears together with the difference between `` entanglement - friendly '' and `` non - friendly '' channels and we recover the eigenvalues obtained in @xcite . again the difference between odd and even dimensions comes form the second term of the expression for @xmath97 in eq . ( [ qcdcev ] ) for even dimensions , which is absent in eq . ( [ qcdcodd ] ) for odd dimensions . and again , due to the factor @xmath149 in front of @xmath97 in eq.([eqcdc ] ) , it is the``entanglement - non - friendly '' phase correlations that are responsible for the observed parity effect . b. schumacher and m. d. westmoreland , phys . rev . a * 56 * , 131 ( 1997 ) ; in _ quantum computation and quantum information : a millenium volume _ , edited by s. lomonaco , contemporary mathematics series ( american mathematical society , providence , rhode island , 2001 ) .
we study the capacity of @xmath0-dimensional quantum channels with memory modeled by correlated noise . we show that , in agreement with previous results on pauli qubit channels , there are situations where maximally entangled input states achieve higher values of mutual information than product states . moreover , a strong dependence of this effect on the nature of the noise correlations as well as on the parity of the space dimension is found . we conjecture that when entanglement gives an advantage in terms of mutual information , maximally entangled states saturate the channel capacity . ``
let @xmath4 be a one - dimensional it semimartingale , which in particular means that its `` continuous martingale part '' has the form @xmath5 where @xmath6 is a standard brownian motion , and the process @xmath7 is optional and ( locally ) squared integrable . one of the long - standing problems is the estimation of the so - called integrated volatility , say at time @xmath8 , that is of the variable @xmath9 , where @xmath10 is the ( squared ) volatility , on the basis of discrete observations of @xmath4 . a huge number of papers have been devoted to this question already , in various situations : when the process is continuous ( so @xmath4 is the sum of @xmath11 above , plus possibly a drift term ) , or when it has jumps ; when the process @xmath4 is `` perfectly '' observed , or contaminated by noise ; when the sampling times are equi - spaced , or when they are irregularly spaced . below , we focus on the basic case , where the sampling is at regularly spaced times @xmath12 for @xmath13 , and when @xmath14 is observed without noise . even in this simple situation , the question of the `` optimal '' rate of convergence of estimators toward @xmath15 , as @xmath16 , is unanswered so far , when there are jumps which are `` too active . '' more precisely , estimators are known , which converge to @xmath15 with the rate @xmath17 , in the continuous case ( the realized volatility , or `` approximate quadratic variation '' at time @xmath8 , achieves this rate ) , and also when @xmath4 has jumps with a degree of activity , or blumenthal getoor index , less than @xmath8 . this rate is optimal ( in a minimax sense ) , for the following reason : if @xmath18 where @xmath19 is a constant , so @xmath20 , we have a purely parametric setting for which the local asymptotic normality ( lan ) holds with rate @xmath17 , and the realized volatility is indeed the mle in this case . however , when the degree @xmath0 of jump activity is larger than @xmath8 , the best rates found in the literature are of the form @xmath21 for @xmath22 arbitrarily small ( see below for more details ) . the difficulty comes of course from the essentially nonparametric feature of the problem , since we do not want to specify the law of the process @xmath4 , apart from the fact that it is an it semimartingale , plus possibly some boundedness assumptions on its characteristics . in a purely parametric problem , for example , when @xmath4 is a lvy process with a _ known _ lvy measure and the only unknown parameters are the variance @xmath23 of the gaussian part , and possibly the drift , then again the rate @xmath17 is available for estimating @xmath23 ( this rate is achieved by the mle , under very general circumstances ) . there has been a considerable interest in providing also nonparametric estimators that converge at rate @xmath17 , but as we show here , this is in general impossible . in this paper , a bound for the minimax rate is determined , when the degree of activity is @xmath0 or smaller [ the precise definition of @xmath0 is given in assumption [ asslr ] below , and is slightly different from the usual blumenthal getoor index ] . we will see that the best possible rate is @xmath24 when @xmath25 ( and of course @xmath17 when @xmath26 ) . it is interesting to notice that the truncated realized volatility , which achieves the rate @xmath21 for any prespecified @xmath22 is indeed `` almost '' rate - optimal . the paper is organized as follows : in section [ sec - k ] , we state the assumptions and review some known results . the results of this paper are presented in section [ sec - r ] , and the proofs are given in the last section . we consider a one - dimensional it semimartingale @xmath4 on a filtered space @xmath27 , which is observed at regularly spaced times @xmath28 for @xmath29 , over the ( fixed ) finite interval @xmath30 $ ] . the characteristics @xmath31 where @xmath32 is the drift , @xmath33 the integrated volatility and @xmath34 the lvy system of @xmath4 ( see , e.g. , chapter 1 of @xcite ) , thus have the form @xmath35 here , @xmath36 and @xmath37 are optional ( or , predictable ) processes , with @xmath38 , and @xmath39 is an optional random measure , also called the lvy measure , which accounts for the jumps of the process . when @xmath4 is continuous , the canonical way for estimating @xmath15 is to use the realized volatility , or approximate quadratic variation at time @xmath8 : @xmath40^n_1=\sum _ { i=1}^n\bigl(\delta^n_ix \bigr)^2\qquad\mbox{where } \delta^n_ix = x_{i / n}-x_{(i-1)/n},\ ] ] which converges in probability to @xmath15 . when further @xmath41 and @xmath42 are a.s . finite , we have the stable convergence in law at rate @xmath17 @xmath43^n_1-c_1 \bigr ) { \stackrel{\mathcal{l}-\mathrm{s}}{\longrightarrow}}\mathcal{u}\qquad\mbox{where } \mathcal{u}=\sqrt{2}\int _ 0 ^ 1c_s \,dw'_s,\ ] ] and where @xmath44 is a standard brownian motion , defined on an extension of @xmath27 , and which is independent of the @xmath45-field @xmath46 : see , for example , theorem 5.4.2 in @xcite . when @xmath4 has jumps , the variables @xmath47^n_1 $ ] no longer converge to @xmath15 , but to the `` full '' quadratic variation @xmath47_1= c_1+\sum_{s\leq1}(\delta x_s)^2 $ ] , where @xmath48 denotes the jump size at time @xmath49 . however , there are two known methods to consistently estimate @xmath15 : _ truncated realized volatility_. one chooses a sequence @xmath50 of positive truncation levels , typically of the form @xmath51 for some @xmath52 , and considers @xmath53 _ multipower variations_. one chooses an integer @xmath54 , and considers @xmath55 where @xmath56 is the @xmath57th absolute moment of a standard normal variable @xmath58 ( other versions are possible ; one may , e.g. , take any product of @xmath59 increments , with powers adding up to @xmath60 ) . the first method has been introduced by mancini in @xcite , the second one by barndorff - nielsen and shephard in @xcite . both provide estimators which converge in probability to @xmath15 , upon rather weak assumptions on the jumps . the question of the rate of convergence , though , is still open , and we quickly review the known results , in the case of truncated realized volatility . one needs the following assumption , where @xmath0 is a number in @xmath61 $ ] : [ asslr ] the variables @xmath62 , @xmath63 and@xmath64 are almost surely finite . the larger @xmath0 is , the weaker assumption [ asslr ] is . ( l-2 ) is a very weak assumption for an it semimartingale , whereas [ asslr ] when @xmath65 puts restrictions on the jump activity , and is slightly stronger than saying that the blumenthal getoor index of @xmath4 ( or , jump activity index ) is not bigger than @xmath0 . in particular , ( l-@xmath8 ) is slightly stronger than the property of the jumps to be summable on each finite interval , for example , the jump part to have trajectories of finite variation . note that a stable process of index @xmath66 satisfies [ asslr ] for all @xmath67 , but not for @xmath68 . when [ asslr ] holds for some @xmath69 , the estimators @xmath70 enjoy exactly the same clt as in ( [ 3 ] ) with @xmath71 in place of @xmath47_t$ ] , with the same limit , provided we have @xmath72 when [ asslr ] holds for some @xmath73 , the clt with rate @xmath17 no longer holds for @xmath71 , but we have when @xmath74 with @xmath75 : @xmath76 ( convergence in probability ) . these results are shown in @xcite , and mancini in @xcite has proved that when the jumps of @xmath4 are those of a stable process with index @xmath77 [ so holds for all @xmath67 , but not for @xmath78 , and when @xmath79 , the estimator converges exactly at rate @xmath80 , in the sense that the sequence @xmath81 converges to a nontrivial limit ( in probability , and not in law , in this case ) : this rate is less than @xmath17 , as it is in ( [ 7 ] ) , and no proper clt is available in this case . turning now to multipowers , we have analogous results , except that one needs stronger assumptions : basically , [ asslr ] plus the fact that the process @xmath37 is also an it semimartingale , and never vanishes : the clt for @xmath82 holds when @xmath69 , with @xmath83 replaced by a suitable ( bigger ) constant depending on @xmath59 ; see @xcite . when @xmath84 , vetter in @xcite proves that there is a clt at rate @xmath17 with a nonvanishing bias term . when @xmath25 nothing is formally known , but the presence of the bias term when @xmath84 suggests that for @xmath25 the rate is less than @xmath17 we are in a nonparametric setting , in which the process @xmath4 is not specified [ apart from the fact that it satisfies [ asslr ] for some @xmath0 ] , and even the space @xmath86 is not specified . the meaning of `` optimality '' or `` rate - optimality '' is not a priori clear ; and , to begin with , even the quantity to estimate , namely @xmath15 , depends of course on the space @xmath86 and on @xmath4 . a possible setting is as follows . we consider a family @xmath87 of it semimartingales satisfying [ asslr ] , each one being defined on its own filtered space @xmath86 , and the quantity to estimate is the associated integrated volatility @xmath88 . each @xmath4 in @xmath87 takes its values , as a process , in the skorokhod space @xmath89 of all cdlg functions on @xmath90 , and the image by @xmath4 of the observed @xmath45-field @xmath91 is the @xmath45-field @xmath92 of @xmath89 . for any @xmath93 we denote by @xmath94 the restriction to @xmath95 of the law of @xmath4 . an estimator at stage @xmath3 is a @xmath95-measurable function @xmath96 on @xmath89 . we say that a sequence @xmath97 of such estimators achieves the uniform rate @xmath98 ( with @xmath99 ) on @xmath87 , for estimating @xmath15 , if the family @xmath100 is tight , uniformly in @xmath3 and in @xmath101 , that is , @xmath102 uniformly in @xmath93 . of course , if @xmath103 denotes the set of _ all _ it semimartingales satisfying [ asslr ] , there can not be any uniform rate because , to begin with , the variables @xmath88 are not uniformly tight when @xmath4 runs through @xmath103 : we need to restrict our attention to subfamilies of @xmath103 which are `` bounded '' in some sense . in view of the formulation of [ asslr ] , it is natural to consider , for any @xmath104 , the class @xmath105\\[-10pt ] & & \mbox{$|b_t|+c_t+\int(|x|^r\wedge1 ) f_t(dx ) \leq a \mbox { for all $ t$}$}.\nonumber\end{aligned}\ ] ] we also denote by @xmath106 the subclass of all lvy processes belonging to @xmath107 . the main result of this paper is the following theorem . [ tr-1 ] let @xmath1 and @xmath104 . any uniform rate @xmath98 for estimating @xmath88 , within the class @xmath106 , hence also within the bigger class @xmath107 , satisfies ( up to a multiplicative constant , of course ) @xmath108 the results recalled in the previous section show that the truncated estimators @xmath70 ( which are estimators in the sense specified above ) achieve the rate @xmath109 when @xmath69 , and at least @xmath110 when @xmath73 , for any @xmath4 satisfying [ asslr ] . we indeed have ( slightly ) more : [ tr-2 ] let @xmath1 and @xmath104 , and take @xmath74 . the truncated estimators @xmath70 have the uniform rate @xmath98 below , within @xmath107 , for estimating @xmath88 , @xmath111 when @xmath69 , the truncated estimators @xmath70 achieve the uniform rate @xmath17 , and as seen in the previous section they even enjoy a clt . when @xmath73 we have the uniform rate @xmath110 , although for any given @xmath4 we indeed have a `` faster '' rate , as seen in ( [ 7 ] ) ; however , this faster rate is not uniform in @xmath112 , as could be seen by taking a sequence of lvy processes with characteristics @xmath113 , with @xmath114 ( so @xmath115 for all @xmath3 ) , but such that @xmath116 does _ not _ tend to @xmath117 as @xmath118 . we then conclude that the truncated estimators are uniformly rate optimal when @xmath69 , and otherwise they approach the bound @xmath109 , up to @xmath119 with @xmath22 arbitrarily small , upon choosing @xmath120 close enough to @xmath121 . let us finally show that on the restricted class @xmath106 of lvy processes the rate @xmath109 of ( [ r-2 ] ) can be achieved exactly and thus constitutes the exact minimax optimal rate : this means that for any @xmath1 and any @xmath104 one can find estimators for @xmath88 enjoying the uniform rate @xmath109 . when @xmath69 , we already know this ( even for the much larger class @xmath107 ) by the previous theorem , but for all @xmath1 we can construct estimators with the uniform rate @xmath109 on @xmath122 as follows . for any process @xmath4 , we consider the empirical characteristic function of the increments , at each stage @xmath3 ( below @xmath123 ) : @xmath124 then we set @xmath125 [ tr-3 ] for all @xmath104 and @xmath1 , the estimators @xmath126 with @xmath127 attain the uniform rate @xmath109 for estimating @xmath88 , within the class @xmath128 of lvy processes . [ r1 ] when @xmath26 the estimators @xmath126 are likely to enjoy a central limit theorem with rate @xmath109 , and with a bias when @xmath84 . when @xmath25 , and upon examining the proof [ see ( [ c-13 ] ) and ( [ c-15 ] ) , e.g. ] , the estimation error @xmath129 is the sum of a random part , which is easily seen to enjoy a clt with rate @xmath130 , and a nonrandom part equal to @xmath131 , where @xmath132 is the lvy measure of the lvy process @xmath4 under consideration . it turns out that @xmath133 tends to @xmath117 by lebesgue s theorem , so , for any given @xmath4 we indeed have in probability : this convergence is of course not uniform in @xmath134 , otherwise the conclusion of theorem [ tr-1 ] would be violated . now , depending on whether @xmath135 converges or diverges and both occurrences are possible we have a clt with rate @xmath136 , or we have a slower effective rate ( still at least @xmath109 , of course ) with the normalized error converging in probability to a nontrivial limit . note that the argumentation is in line with the standard nonparametric error decomposition in a bias and variance part . our estimator uses that the characteristic exponent for high frequencies @xmath137 separates the brownian from the jump part according to the ratio @xmath138 . we have reliable empirical access to this exponent only up to frequency @xmath137 ( otherwise the stochastic error explodes due to a gaussian deconvolution setting ) . so far , we do not know whether this spectral approach yields the same optimal rate on the larger class @xmath107 . _ the bound _ @xmath139 . for proving this bound , it is enough to show that it already holds on the subclass @xmath140 of all brownian motions with unit variance @xmath141 ( so @xmath142 for all @xmath143 $ ] ) . in this case , and as already mentioned in the , the increments follow the parametric model @xmath144 with parameter @xmath23 running through @xmath145 $ ] , for which the lan property holds with rate @xmath17 , and the result follows . _ the bound _ @xmath146 _ when _ @xmath147 . by scaling , if the result holds for one @xmath104 , it holds for all @xmath104 . hence , in order to find a bound on the uniform rate @xmath98 on @xmath122 , hence a fortiori on @xmath107 , it is enough to construct two sequences @xmath148 and @xmath149 of lvy processes belonging to @xmath150 for @xmath151 and some constant @xmath152 , with the following two properties , where @xmath153 : @xmath154 indeed , letting @xmath155 be a sequence of estimators with uniform rate on @xmath107 ( or , even , on @xmath156 ) , the two sequences @xmath157 and @xmath158 are tight under @xmath159 and @xmath160 , respectively , by ( [ p-1 ] ) . then ( [ p-2 ] ) implies that the sequence @xmath161 is also tight under @xmath160 . this is possible only if the sequence @xmath162 is bounded . so @xmath163 is an upper bound for any uniform rate on @xmath164 ( up to a multiplicative constant , of course ) . we take lvy processes @xmath148 and @xmath149 with respective characteristics @xmath165 and @xmath113 , with lvy measures @xmath166 satisfying @xmath167 for some constant @xmath152 ( below constants change from line to line , and may depend on @xmath0 , and are all denoted as @xmath152 ) . by construction , we have ( [ p-1 ] ) and @xmath168 for a constant @xmath152 [ which may differ from the one in ( [ c-1 ] ) ] , and we need to choose the above measures @xmath169 and @xmath170 in such a way that ( [ p-2 ] ) is satisfied . we use the following convention for the fourier transform , namely @xmath175 , so the inverse is @xmath176 . we also denote as @xmath177 the @xmath178th derivative of any @xmath178-differentiable function @xmath179 . since @xmath180 for all @xmath181 and @xmath182 , we can define @xmath183 , and we have @xmath184 , where @xmath185 . by the plancherel identity we deduce @xmath186\\[-8pt ] & & \qquad \rightarrow\quad \|h_{n , q } \|_{\mathbb{l}^2}\leq\bigl\|h_{n}^{(q)}\bigr\|_{\mathbb{l}^2}\leq ka_n.\nonumber\end{aligned}\ ] ] then the cauchy schwarz inequality applied to the functions @xmath187 and @xmath188 yields @xmath189 [ note that @xmath190 is bounded in @xmath3 when @xmath191 , but not otherwise ; we also have @xmath192 . therefore , the two measures @xmath193 are nonnegative and integrate @xmath194 , hence are lvy measures . this construction will satisfy ( [ p-2 ] ) mainly because the definition of the two lvy measures and the constant value of @xmath195 for @xmath196 imply that the difference between the two characteristic exponents vanishes for @xmath196 , as we shall prove next . splitting the integration domain into the sets @xmath197 and @xmath198 in the integral @xmath199 , we get @xmath200 in turn , the integral @xmath201 can be split into integrals on the sets @xmath202 , @xmath203 and @xmath204 , and recalling @xmath205 we deduce from the above that @xmath206 it follows that the measures @xmath169 and @xmath170 satisfy ( [ c-1 ] ) , and it remains to prove ( [ p-2 ] ) . we denote by @xmath207 and @xmath208 the characteristic functions of @xmath209 and @xmath210 , and @xmath211 . these functions are real ( because @xmath212 is an even function ) and are given by @xmath213 where @xmath214 we proceed to studying @xmath215 and @xmath216 . equation ( [ c-3 ] ) applied with @xmath217 implies that @xmath215 and @xmath216 are twice differentiable . first , we have @xmath218 , hence ( [ c-4 ] ) yields @xmath219\\[-8pt ] \bigl|{\widetilde{\phi}}'_n(u)\bigr|&\leq & k \bigl(1+a_n^{(3 - 2r)/(4 - 2r)}\bigr ) |u|.\nonumber\end{aligned}\ ] ] second , @xmath220 , whereas @xmath221 , and this yields @xmath222 \\[-8pt ] since @xmath148 and @xmath149 have a nonvanishing gaussian part , the variables @xmath209 and @xmath210 have densities , denoted by @xmath223 and @xmath224 , and we set @xmath225 . since @xmath148 and @xmath149 are lvy processes , the variation distance between @xmath159 and @xmath226 is not more than @xmath3 times @xmath227 , and we are thus left to show that @xmath228 . to check this , we use the same argument as for ( [ c-4 ] ) : if @xmath229 , by the cauchy schwarz inequality we have @xmath230 , whereas @xmath231 and also , since @xmath232 is twice differentiable , @xmath233 . by plancherel identity , it is thus enough to prove that @xmath234 we have @xmath235 and @xmath236 , which implies @xmath237 and @xmath238 , whereas @xmath239 if @xmath196 and @xmath240 if @xmath241 by ( [ c-6 ] ) . therefore , @xmath242 and also , upon using ( [ c-5 ] ) , @xmath243 now , since @xmath244 , we have @xmath245 for @xmath246 . since further @xmath247 , we deduce @xmath248 then ( [ c-7 ] ) follows , and the proof is complete . any @xmath112 can be written as follows , on some space @xmath27 : @xmath249 here , @xmath250 and @xmath23 are as in [ asslr ] , and @xmath6 is a standard brownian motion , and @xmath251 is a poisson random measure on @xmath252 with intensity measure @xmath253 , and @xmath254 is a predictable function on @xmath255 . the connection between @xmath256 and @xmath257 is that @xmath258 is the image of lebesgue measure by the map @xmath259 , restricted to @xmath260 . with @xmath98 given by ( [ r-3 ] ) , it is clearly enough to prove that , for some constant @xmath152 only depending on @xmath265 ( as will be all constants @xmath152 below , changing from line to line ) , we have @xmath266 here , we recall estimates on the increments of @xmath267 and @xmath263 , the later coming from lemmas 2.1.5 and 2.1.6 of @xcite , and where @xmath268 is arbitrary ( the constants @xmath269 below depend on @xmath57 in addition to @xmath270 ) . namely , since @xmath271 , we have uniformly in @xmath272 $ ] : @xmath273\\[-8pt ] \mathbb{e}\bigl(\bigl|y_s - y_{(i-1)/n}\bigr|^p \bigr)&\leq & kn^{-(p / r)\wedge1}.\nonumber\end{aligned}\ ] ] we will also use the following estimates , which follow from the property @xmath274 and from the fact that if @xmath275 there is at least one jump of @xmath264 within the interval @xmath276 $ ] ( this estimate follows from lemma 2.1.7 of @xcite applied to the counting process @xmath277 ) : @xmath278 with the notation ( [ 2 ] ) , it s formula yields @xmath279^n_1-c_1=u_n+v_n$ ] , where @xmath280 equation ( [ c-20 ] ) yields @xmath281 whereas @xmath282 . thus we have @xmath283 and @xmath284 , implying @xmath285^n_1-c_1\bigr| \bigr)\leq k/\sqrt{n}.\ ] ] therefore , it remains to prove that @xmath286^n_1\bigr| \bigr ) \leq k / w_n.\ ] ] consider the case @xmath69 first . by lemma 13.2.6 of @xcite , applied with @xmath287 and @xmath288 , hence @xmath289 and @xmath290 and @xmath291 ( with the notation of this lemma ) , we have @xmath292 where the last inequality follows from @xmath293 . on the other hand , ( [ c-20 ] ) and markov inequality yield @xmath294 for any , and upon taking @xmath295 we obtain @xmath296^n_1- \sum_{i=1}^n\bigl(\delta^n_ix ' \bigr)^2 1_{\ { now we turn to the case @xmath73 . one has @xmath297^n_1 = \break\sum_{j=1}^3u(j)_n$ ] , where @xmath298 and @xmath299 it s formula yields , with the notation @xmath300 , and taking advantage of the facts that @xmath263 and @xmath301 are two orthogonal martingales and that @xmath302 is a martingale : @xmath303 then standard estimates and ( [ c-20 ] ) , plus hlder s inequality , yield ( the first bound is a.s . ) @xmath304 since @xmath305 , these estimates yield @xmath306 and@xmath307 , hence it is enough to show that @xmath308 . recalling @xmath73 , the following inequality is easy to check , for @xmath309 and @xmath310 $ ] : @xmath311 it follows that @xmath312 , where @xmath313 equation ( [ c-21 ] ) yields @xmath314 , and ( [ c-20 ] ) yields @xmath315 . another application of ( [ c-20 ] ) , plus hlder and markov inequalities , give us @xmath316 for @xmath317 . upon taking @xmath57 large enough , we obtain @xmath318 for @xmath319 . we deduce @xmath308 , and the proof is complete . we let @xmath320 , where @xmath1 and @xmath104 are given . the characteristic triple of @xmath4 is @xmath321 and the characteristic function of @xmath322 is @xmath323 then @xmath324 , where @xmath325 . as soon as @xmath3 is large enough we have @xmath326 , hence , since @xmath327 , @xmath328 because @xmath329 by hypothesis , and in view of the form of @xmath137 in ( [ r-30 ] ) , by singling out the two cases @xmath26 and @xmath25 this implies that , with @xmath330 , @xmath331 second , we study @xmath212 . the variables @xmath337 are i.i.d . as @xmath338 varies , with modulus @xmath8 and expectation @xmath339 , hence @xmath340 satisfies @xmath341 . in view of ( [ c-13 ] ) , on the set @xmath342 we have @xmath343 and @xmath344 as soon as @xmath345 , in which case we deduce , for some universal constant @xmath152 : @xmath346 henceforth , if @xmath347 , @xmath348 putting together ( [ c-14 ] ) and ( [ c-15 ] ) , plus the fact that @xmath349 ( by bienaym ' e tchebycheff inequality ) tends to zero , and the equality @xmath350 , we deduce that @xmath351 [ with the notation ( [ r-2 ] ) ] is tight , uniformly in @xmath320 .
the optimal rate of convergence of estimators of the integrated volatility , for a discontinuous it semimartingale sampled at regularly spaced times and over a fixed time interval , has been a long - standing problem , at least when the jumps are not summable . in this paper , we study this optimal rate , in the minimax sense and for appropriate `` bounded '' nonparametric classes of semimartingales . we show that , if the @xmath0th powers of the jumps are summable for some @xmath1 , the minimax rate is equal to @xmath2 , where @xmath3 is the number of observations .
the observation of the sky with space - born instruments , equipped with detectors working at those electro - magnetic frequencies that can not be accessed from the ground , revealed the existence of several classes of high energy radiation sources . with their location in distant galaxies , active galactic nuclei ( agns , in brief ) turned out to be the most powerful non transient sources of such radiation . although agns appear with many different properties , they share an extreme intrinsic luminosity , ranging between @xmath5 and @xmath6 , comparable to or greater than the energy output of large galaxies , but released from a region that is smaller than 1 pc in radius . to explain this property we assume that large amounts of matter , in the order of some solar masses per year , are conveyed to the nuclear regions of active galaxies , where they are accreted by a super massive black hole ( smbh ) . it is now well established that smbhs with masses between @xmath7 and some @xmath8 reside in the nuclei of every massive galaxy ( ferrarese & merrit , 2000 ; shankar , 2009 ) . in addition , we know that accretion of fuel into their gravitational field leads to the conversion of gravitational binding energy into radiation with very high efficiency ( blandford & znajek , 1977 ; shields , 1978 ) . due to the presence of relativistic plasmas and strong magnetic fields in the vicinity of the black hole s accretion flow , the spectrum of the emitted radiation results from a combination of thermal and non - thermal components that cover several orders of magnitude in frequency , sometimes extending from radio wavelengths all the way up to @xmath0-ray energies . the fermi large area telescope ( fermi - lat ; see atwood et al . , 2009 ) gave new life to the study of the @xmath0-ray sky , producing , after 4 years of scientific observations , a map of @xmath0-ray detections with unprecedented resolution and sensitivity in the energy range between 100 mev and 300 gev ( acero et al . thanks to this result , a large number of @xmath0-ray sources can now be associated with lower energy counterparts . in the extra - galactic environment , it turned out that the most commonly detected objects are agns belonging to the blazar class . blazars are extremely variable , highly polarized , radio - loud sources , dominated by power - law continuum spectra of type @xmath9 , with a typical radio spectral slope @xmath10 . their properties are the consequence of the relativistic beaming of the synchrotron radiation produced by a jet that is collimated and accelerated close to our line of sight ( blandford & knigl , 1979 ) . they are classically separated into bl lac objects ( bll ) , whose optical spectra show a nearly featureless power - law continuum , and flat spectrum radio quasars ( fsrqs ) , which , instead , are characterized by strong emission lines . in addition to the blazars that dominate the extra - galactic @xmath0-ray population , other types of agns , together with a large number of sources without a firm classification , are detected as well . in this contribution we describe our investigation on the nature of @xmath0-ray agns of undetermined type , through the observation of their optical spectra . we focus our attention on targets whose spectral energy distributions ( sed ) are consistent with those of blazars , although they still lack firm spectroscopic classification , and they are therefore called blazar candidates of undetermined type ( bcu in 3fgl terminology , acero et al . , 2015 ) . in this report , we discuss the spectral classification of some bcus , with respect to the associated seds , rather than giving a full list of observations . in the following sections , we describe the techniques used in the attempt to identify the low energy counterparts of the @xmath0-ray sources , the details of how we collected the optical spectra and their classification criteria . finally , we draw a sketch of the @xmath0-ray emission in the different classes of objects that we observe , compared with the multiple wavelength properties of their seds . at the energies of @xmath0-ray photons it is not possible to focus radiation through reflecting or refracting optical devices . the fermi - lat , instead , measures the production of @xmath11 pairs , through the conversion of photons in the detector ( atwood et al . , 2009 ) . the pair properties are used to reconstruct the energy and the direction of the incoming photon . the precision that can be achieved in the measurement depends mainly on the incoming photon incidence angle and on its energy . it can be roughly estimated that , for a photon of energy between 10 gev and 100 gev , hitting the detector at normal incidence , the 95% containment angle is approximately 0.5@xmath12(ackermann et al . , 2012 ) . although the detection of multiple photons from a source can improve the performances , the contribution of nearby sources and background noise , combined with less than optimal detection conditions , results in a still significant uncertainty in the localization of the sources . thus , in general , the task to associate a low energy counterpart to @xmath0-ray sources is not trivial . in the case of agns , the expected broad band emission provides a reliable way to better constrain the source position . since @xmath0-rays from agns arise from inverse compton ( ic ) scattering of low energy seed photons by relativistic plasma particles confined in strong magnetic fields , the @xmath0-ray production occurs together with synchrotron radiation . if the energy distribution of the plasma particles is extended enough to support significant @xmath0-ray emission , we expect that powerful synchrotron radiation is produced at radio and x - ray energies , as well . taking advantage from the angular resolution of instruments working at these frequencies , which can measure the position of radiation sources down to a few arcseconds , we are able to identify candidate counterparts to @xmath0-ray emission by looking for coincident x - ray and radio sources within the @xmath0-ray detection uncertainty radius ( ackermann et al . , 2011 ; gasparrini et al . furthermore , the existence of a connection between @xmath0-ray emission and lower frequency radiation implies that large flux variations , that characterize agns particularly in the high energies , can be observed in different frequencies . thanks to the monitoring strategy of fermi - lat observations , covering the entire sky approximately every 3 hr , the detection of important flaring activity can be matched with follow - up observations that are able to identify the possible correlated variations of the source in other frequencies . once the source position is determined by either of the techniques down to a few arcseconds , it is possible to search for the optical counterpart and to obtain its spectrum . the latest catalog of fermi - lat detected agns ( 3lac , see ackermann et al . , 2015 ) includes blazars , radio galaxies , steep spectrum radio quasars ( ssrqs ) , seyfert galaxies and narrow line seyfert 1 galaxies . in addition to the classified sources , however , a number of undetermined type objects , generally called agns or bcus , still exists . focusing our attention on these unclassified objects , we carried out a spectroscopic study that combines publicly available spectra , like those extracted from the latest data release of the sloan digital sky survey ( sdss - dr 12 ; alam et al . , 2015 ) and of the 6df galaxy redshift survey ( 6dfgrs - dr 3 ; jones et al . , 2004 , 2009 ) , together with new observations performed at the 1.22 m and the 1.82 m telescopes of the asiago astrophysical observatory ( ciroi et al . , 2014 ) . 1 illustrates examples of spectra obtained in this study . the main purpose of this optical spectroscopic analysis is to identify the agn class for the source associated to the @xmath0-ray emission and to determine its redshift , through the identification of known emission or absorption lines . in general , the detection of emission lines ensures the most reliable measurement of the source redshift , because they are easier to detect and they are originated very close to central smbh . absorption lines , on the contrary , are produced either in the host galaxy or along the light path towards us . they are much more difficult to detect and can only be seen if the agn is not overwhelmingly dominant . it follows that the presence of emission lines characterizes agns with radiatively efficient accretion activity , therefore surrounded by nuclear and circum - nuclear regions of ionized gas , like quasars and seyfert galaxies . blls , on the contrary , just give raise to a featureless power - law continuum , where some absorption features can show up , if the jet power is relatively low with respect to the host galaxy luminosity , or when some intervening medium happens to lie along the line of sight to the source . .list of sources included in this report . the table provides the @xmath0-ray source identifier , the associated counterpart name , the counterpart position ( j2000.0 ) , the object redshift and its spectroscopic classification . [ cols="<,<,^,^,^,>",options="header " , ] @xmath13 residuals of ssc models are computed without taking into account the thermal excess data points . in this study we presented the optical spectra of six targets that have been associated to @xmath0-ray sources of still undetermined type in 3lac . in some cases , the spectra of this type of objects can still be obtained from large spectroscopic surveys , such as the sdss , above all , but , in others , we need specifically planned observations . the increase of sensitivity and resolution achieved by the fermi - lat at @xmath0-ray energies is now providing better opportunities to identify candidate counterparts to extra - galactic @xmath0-ray emission and , therefore , to improve the selection of targets . with an improved ability to detect @xmath0-rays from faint sources we can now investigate the occurrence of nuclear activity on different power scales . the detection of faint blazar - like activity in low luminosity agns or even apparently normal galaxies opens a new window on the demographics of @xmath0-ray sources , as well as on the mechanisms that contribute to black hole growth and jet formation . the possibility that such objects may represent an important contribution to the @xmath0-ray radiation of undetermined origin deserves further investigation . searching for radiation from hidden agns is a fundamental science case for instruments designed to observe high energy photons and other hints of jet activity , such as light polarization . therefore , we plan to further investigate the spectroscopic properties of candidate counterparts to @xmath0-ray emission , taking possibly into account light polarization studies as well , through an extensive observational campaign designed for middle class telescopes . the _ fermi_-lat collaboration acknowledges support for lat development , operation and data analysis from nasa and doe ( united states ) , cea / irfu and in2p3/cnrs ( france ) , asi and infn ( italy ) , mext , kek , and jaxa ( japan ) , and the k.a . wallenberg foundation , the swedish research council and the national space board ( sweden ) . science analysis support in the operations phase from inaf ( italy ) and cnes ( france ) is also gratefully acknowledged . this work is based on observations collected at copernico ( or / and schmidt ) telescope(s ) ( asiago , italy ) of the inaf - osservatorio astronomico di padova . part of this work is based on archival data , software or online services provided by the asi science data center ( asdc ) .
during its first four years of scientific observations , the fermi large area telescope ( fermi - lat ) detected 3033 @xmath0-ray sources above a 4@xmath1 significance level . although most of the extra - galactic sources are active galactic nuclei ( agn ) of the blazar class , other families of agns are observed too , while a still high fraction of detections ( @xmath2 ) remains with uncertain association or classification . according to the currently accepted interpretation , the agn @xmath0-ray emission arises from inverse compton ( ic ) scattering of low energy photons by relativistic particles confined in a jet that , in the case of blazars , is oriented very close to our line of sight . taking advantage of data from radio and x - ray wavelengths , which we expect to be produced together with @xmath0-rays , providing a much better source localization potential , we focused our attention on a sample of @xmath0-ray blazar candidates of undetermined type ( bcus ) , starting a campaign of optical spectroscopic observations . the main aims of our investigation include a census of the agn families that contribute to @xmath0-ray emission and a study of their redshift distribution , with the subsequent implications on the intrinsic source power . we furthermore analyze which @xmath0-ray properties can better constrain the nature of the source , thus helping in the study of objects not yet associated with a reliable low frequency counterpart . in this communication we report on the instruments and techniques used to identify the optical counterparts of @xmath0-ray sources , we give an overview on the status of our work , and we discuss the implications of a large scale study of @xmath0-ray emitting agns . @xmath3_dipartimento di fisica e astronomia `` g. galilei '' , universit degli studi di padova , + vicolo dellosservatorio 3 , 35122 - padova ( italy ) _ + @xmath4_iasf milano - istituto di astrofisica spaziale e fisica cosmica , + via e. bassini 15 , 20133 - milano ( italy ) _ _ keywords _ : catalogues galaxies : active galaxies : jets quasars : emission lines
the recent paper by dunkel and hilbert titled `` consistent thermostatistics forbids negative absolute temperatures '' @xcite has triggered a vigorous debate on whether the boltzmann entropy ( alias the surface entropy , eq . [ eq : somega ] ) or the gibbs entropy ( alias the volume entropy , eq . [ eq : somega ] ) is the more appropriate expression for the thermodynamic entropy of thermally isolated mechanical systems @xcite . the thermodynamic consistency of the gibbs entropy has been a leitmotiv that sporadically recurred in the classical statistical mechanics literature . it started with helmholtz @xcite , boltzmann @xcite , and gibbs @xcite , it continued with p. hertz @xcite einstein @xcite , and others @xcite , until it has been reprised recently by various authors @xcite . this line of research culminated with the work of ref . @xcite , showing that the gibbs entropy complies with all known thermodynamic laws and unveiling the mistakes apparently incurred into the arguments of its opponents @xcite . while the work of ref . @xcite is characterised by a top - down approach ( namely , one postulates an entropy expression and then investigates compliance with the thermodynamic laws ) here we adopt instead a bottom - up approach : we begin from the thermodynamic laws and construct the expression of the microcanonical entropy on them . in particular we base our construction on the following two fundamental pillars of thermodynamics . 1 ) the second law of thermodynamics as formulated by clausius for quasi - static processes , namely , @xmath4 , which says that @xmath5 is an integrating factor for @xmath6 , and identifies the entropy with the associated primitive function @xmath7 . 2 ) the equation of state of an ideal gas @xmath8 . our construction , based on the mathematics of differential forms , leads uniquely to the gibbs entropy ; see sec . [ sec : construction ] . as a consequence the adoption of any expression of entropy other than the gibbs entropy , e.g. , the boltzmann entropy , may lead to inconsistency with the fundamental pillars . this will be illustrated with a macroscopic collection of spins in a magnetic field . as we will see the boltzmann entropy severely fails to predict the correct value of the magnetization , and even predicts a nonexistent phase transition in the thermodynamic limit , see sec . [ subsec : failure ] . this provides a compelling reason for discarding the boltzmann entropy at once . the present work thus complements the work of ref . @xcite by stating not only the compliance of the gibbs entropy with the thermodynamic laws , but also its _ necessity _ and _ uniqueness _ : thermodynamic entropy has to be expressed by means of gibbs formula and no other expression is admissible . together with ref . @xcite the present work appears to settle the debated issue . we recall the definitions of boltzmann and gibbs entropies within the microcanonical formalism @xcite : @xmath9\label{eq : somega}\ , , \\ s_g(e,\bm{\lambda } ) & = k_b \ln \omega(e,\bm{\lambda } ) \label{eq : somega } \ , , \end{aligned}\ ] ] where @xmath10 \label{eq : phi}\end{aligned}\ ] ] denotes the volume of the region of the phase space of the system with energy not above @xmath11 . the symbol @xmath12 stand for some arbitrary constant with units of energy . here @xmath13 , denotes the hamilton function of either a classical or a quantum system with degrees of freedom @xmath14 and @xmath15 denotes external parameters , e.g. the volume of a vessel containing the system or the value of an applied magnetic or electric field @xcite . @xmath16 is their number . in the case of continuous classical systems the symbol @xmath17 stands for an integral over the phase space normalized by the appropriate power of planck s constant and possible symmetry factors . for classical discrete systems , @xmath17 denotes a sum over the discrete state space . for quantum systems @xmath17 is the trace over the hilbert space . the symbol @xmath18 stands for the heaviside step function . the symbol @xmath19 stands for the density of states , namely the derivative of @xmath20 with respect to @xmath11 : @xmath21 = \frac{\partial \omega(e,\bm{\lambda } ) } { \partial e}\ , . \label{eq : omega}\end{aligned}\ ] ] here it is assumed that the spectrum is so dense that the density of states can be considered a smooth function of @xmath11 . the main objective is to link thermodynamic observables , i.e. the forces @xmath22 and temperature @xmath23 , to the quantities which naturally pertain to both the mechanical hamiltonian description and the thermodynamic description , i.e. , the energy @xmath11 and the external parameters @xmath24 . as we will see the entropy , @xmath7 will follow automatically and uniquely once the @xmath22 s and @xmath23 are linked . we begin with the thermodynamic forces @xmath22 , whose expression is universally agreed upon @xcite : @xmath25 with @xmath26 denoting the ensemble average . within the microcanonical framework these are expressed as : @xmath27 } { \omega(e,\bm{\lambda } ) } \right ) \ , . \label{eq : fmicro}\end{aligned}\ ] ] with the expression of @xmath28 we can construct the differential form representing heat : @xmath29 @xmath6 is a differential form in the @xmath30 dimensional space @xmath31 . it is easy to see that , in general @xmath6 is not an exact differential ; see , e.g. , ref . @xcite . before we proceed it is important to explain the meaning of @xmath32 within the microcanonical formalism . the idea behind the microcanonical ensemble is that @xmath11 and @xmath24 are controllable parameters . and @xmath24 are controllable parameters in the canonical formalism ] accordingly , if the system is on an energy surface identified by @xmath31 , the idea is that the experimentalist is able to steer it onto a nearby energy shell @xmath33 . in practice this is can be a difficult task . it can be accomplished , in principle , in the following way : the experimentalist should first change the parameters by @xmath34 in a quasi - static way . this induces a well defined energy change @xmath35 , which is the work done on the system . this brings the system to the energy shell @xmath36 . to bring the system to the target shell @xmath33 the experimentalist must now provide the energy @xmath37 by other means while keeping the @xmath24 fixed . for example she can shine targeted amounts of light on the system , from a light source . after the energy @xmath37 is absorbed by the system ( or emitted , depending on its sign ) , no other interaction occurs and the system continues undisturbed to explore the target shell @xmath33 . in this framework the light source acts as a reservoir of energy , and the quantity @xmath38 , identified as heat , represents the energy it exchanges . according to the second law of thermodynamics in the formulation given by clausius the inverse temperature @xmath5 is an integrating factor for @xmath6 . this fundamental statement is often called the _ heat theorem _ @xcite . we recall that an integrating factor is a function @xmath39 such that @xmath40 equals the total differential @xmath41 of some function @xmath42 called the associated primitive , or in brief , just the primitive . primitives are determined up to an unimportant constant , which we will disregard in the following . entropy is defined in thermodynamics as the primitive associated with clausius s integrating factor @xmath5 @xcite : @xmath43 in searching for thermodynamically consistent expressions of temperature within the microcanonical formalism , one should therefore look among the integrating factors of the microcanonically calculated heat differential in ( [ eq : deltaq ] ) . it must be remarked that it is not obvious that one integrating factor exists , because the existence of integrating factors is not guaranteed in spaces of dimensions higher than @xmath44 . so the existence of a mechanical expression for thermodynamic temperature ( hence of the entropy ) is likewise not obvious . it turns out however that an integrating factor for the differential in ( [ eq : deltaq ] ) always exists . finding it is straightforward if one re - writes the forces in the following equivalent form @xmath45 } { \omega(e,\bm{\lambda } ) } \right ) \nonumber\\ & = \frac{1}{\omega(e,\bm{\lambda } ) } \tr\ , \left ( \frac{\partial \theta[e - h(\bm{\xi};\bm{\lambda})]}{\partial \lambda_i}\right ) \nonumber \\ & = \frac{1}{\omega(e,\bm{\lambda } ) } \frac{\partial \omega(e,\bm{\lambda})}{\partial \lambda_i}\ , . \label{eq : fomega}\end{aligned}\ ] ] this follows from the fact that dirac s delta is the derivative of heaviside s step function . with this , eq . ( [ eq : deltaq ] ) reads @xmath46 it is now evident that @xmath47 is an integrating factor : @xmath48 @xmath2 being the associated primitive . this does not mean that @xmath49 should be identified with temperature and accordingly @xmath2 with entropy . in fact if an integrating factor exists , this identifies a whole family of infinitely many integrating factors . to find the family of integrating factors , consider any differentiable function @xmath50 with non null derivative @xmath51 . its total differential reads : @xmath52 \delta q\ , . \label{eq : dg}\end{aligned}\ ] ] this means that any function @xmath39 of the form @xmath53 is an integrating factor for the heat differential @xmath6 , and @xmath50 is the associated primitive . in fact all integrating factors must be of the form in eq . ( [ eq : beta ] ) , which is equivalent to saying that all associated primitives must be of the form @xmath54 to prove that all primitives must be of the form in eq . ( [ eq : g(phi ) ] ) we consider the adiabatic manifolds , namely the @xmath16 dimensional manifolds in the space @xmath31 identified by the condition that @xmath55 , i.e. , @xmath56 . note that the density of states is a strictly positive function @xmath57 . this is because increasing the energy results in a strictly larger enclosed volume in the phase space . thus , the adiabatic manifolds are characterised by the condition @xmath58 , ( i.e. , any path occurring on them involves no heat exchanges ) , and each value of @xmath2 identifies one and only one adiabatic manifold . any primitive @xmath42 associated with an integrating factor @xmath59 stays constant on the adiabatic manifolds : @xmath60 unless @xmath59 diverges , which we exclude here . hence the only way by which any primitives @xmath42 and @xmath20 can both be constant on all adiabatic manifolds is that @xmath61 is a function of @xmath2 , as anticipated . note that this rules out automatically the surface entropy @xmath62 $ ] because , in general , the density of states can not be written as a function of the phase volume @xmath20 . this is clear for example in the case of an ideal monoatomic gas in a vessel of volume @xmath63 , for which @xmath64 and @xmath65 @xcite ; see below . our derivation above tells us that the second law requires that the entropy , which is one of the primitives , has to be a function @xmath50 of the phase volume , but does not tell us which function that is . for that we need to identify which , among the infinitely many integrating factors , @xmath66 corresponds to clausius s notion of temperature . we remark that once the function @xmath67 is chosen , it has to be one and the same for all systems . this is because by adjusting the external parameters @xmath24 , whose number and physical meaning is completely unspecified , one can transform any hamiltonian into any other . this fact reflects the very essence of clausius s heat theorem , namely , that there exists a unique and universal scale of temperature which is one and the same for all systems @xcite . we proceed then to single out the function @xmath67 that is consistent with the notion of temperature of an ideal monoatomic gas in a vessel of volume @xmath63 , taking its equation of state @xmath68 as the definition . the hamilton function of an ideal monoatomic gas reads @xmath69 with @xmath70 representing the box potential confining the gas within the volume @xmath63 . the phase volume reads @xcite @xmath71 hence , using eq . ( [ eq : fomega ] ) , we obtain for the pressure , @xmath72 : @xmath73 . confronting this with the ideal gas law we obtain @xmath74 consistently with what is known from thermodynamics . since @xmath75 in this case , we readily recognize that @xmath76 , namely , @xmath77 that is @xmath78 , which singles out the gibbs entropy , @xmath79 as the primitive associated with the integrating factor corresponding to the thermodynamic absolute temperature @xcite . in sum : if one accepts the microcanonical expression ( [ eq : f ] ) of the forces , the gibbs entropy represents the _ only _ expression of thermodynamic entropy which is consistent with the second law of thermodynamics , eq . ( [ eq : deltaq ] ) , and the equation of state of the ideal gas . as mentioned above the density of states is definite positive @xmath80 , also , by definition , the volume @xmath2 is non - negative . hence their ratio @xmath81 is non - negative . this means that , within the _ microcanonical _ formalism , negative temperatures are inadmissible . often the present microcanonical scenario is confused with the more common canonical scenario , where the system stays in a canonical state at all times during a transformation , e.g. , ref . this is unfortunate because , as we see below , microcanonical and canonical descriptions are not equivalent for those finite spectrum systems usually discussed in this context . the same construction presented above can be repeated for systems obeying statistics other than microcanonical @xcite . if applied to the canonical ensemble , @xmath82 , ( with @xmath83 being the canonical partition function ) the canonical expression @xmath84 for the forces , along with the equation of state of the ideal gas , _ uniquely _ identifies the canonical parameter @xmath59 as the integrating factor , and its associated primitive @xmath85 as the only thermodynamically consistent expressions of inverse temperature and entropy within the canonical formalism . , that is , the canonical entropy coincides with the gibbs - von neumann information of the canonical distribution @xmath86 . ] in the canonical formalism nothing formally constraints the sign of @xmath59 to be definite . a spin system in a canonical state at negative @xmath59 will have a positive internal energy @xmath87 . the same system in the microcanonical state of energy @xmath88 will , however , have a positive thermodynamic temperature . this evidences the inequivalence of canonical and microcanonical ensembles in systems with a finite spectrum . in an attempt to justify the correctness of the boltzmann entropy , frenkel and warren @xcite , provided a construction which leads to the boltzmann entropy . it must be stressed that the construction presented by frenkel and warren @xcite is approximate , and valid only under the assumption that the saddle point approximation holds . this approximation holds only when the density of states increases exponentially with the energy . under this assumption , however , the density of states @xmath47 and phase volume @xmath2 coincide . so the construction of frenkel and warren @xcite can not shed light onto which entropy expression is appropriate in the case when they do not coincide , which is indeed the very case of practical interest . in contrast , the present construction is exact , i.e. , it holds regardless of the functional dependence of the density of states on energy . accordingly it says that in any case , independent of whether equivalence of the two entropies holds , the volume entropy is the consistent choice . for continuous classical hamiltonian systems , thanks to the equipartition theorem @xcite , the thermodynamic temperature @xmath23 is identical with the equipartition temperature @xmath89 : @xmath90 where the average is the microcanonical average on the shell @xmath31 . this provides further evidence that the choice @xmath91 conforms to the common notion of temperature of any classical system , not just the ideal monoatomic gas . we further remark that the equipartition theorem also identifies the temperature @xmath92 in eq . ( [ eq : partiallogphi ] ) as an intensive quantity , namely a property that is equally shared by all subsystems @xcite . we emphasize that at variance with previous approaches to the foundations of the gibbs entropy @xcite , which postulated that the thermodynamic temperature is the equipartition temperature , here we have instead postulated only that temperature is the integrating factor that is consistent with the ideal gas law and have obtained the coincidence with the equipartition temperature as an aftermath . the advantage of the present approach is evident : it applies to any microcanonical system , even those for which there is no equipartition theorem ( e.g. , quantum systems ) . at variance with other approaches we chose as starting point the expression for the microcanonical forces ( [ eq : fmicro ] ) which is universally agreed upon and built our construction on that firm ground . the salient point of our argument is the identity ( [ eq : fomega ] ) expressing the microcanonical forces in terms of the partial derivatives of @xmath2 . the identity ( [ eq : fomega ] ) alone has as a consequence that the entropy must be of the form @xmath93 with some @xmath67 with non null derivative @xmath51 . in fact , for any @xmath94 one finds the forces @xmath95 , to be identical to the microcanonical forces @xmath22 , eq . ( [ eq : fmicro ] ) @xmath96 here @xmath97 is a shorthand notation for @xmath98 . if one employs an entropy expression that is not of the form @xmath50 , e.g. , the boltzmann entropy , one can well end up in wrongly evaluating the forces . this happens , for example , in the case of a large collection of @xmath99 non interacting @xmath100 spins in a magnetic field @xmath101 , at energy @xmath11 @xcite , that is the prototypical example of the emergence of negative boltzmann temperature @xcite . the hamiltonian reads @xcite @xmath102 here @xmath101 plays the role of the external parameter @xmath103 , @xmath104 is @xmath105 depending on whether the spin points parallel ( up ) or antiparallel ( down ) to the field , and @xmath106 is the magnetic moment of each spin . at energy @xmath11 , the magnetization is given by ( [ eq : fmicro ] ) : @xmath107 the number of states with @xmath108 spins up is @xmath109 the number of states with no more than @xmath108 spins up is @xmath110 using the relation @xmath111 , and treating @xmath11 as a continuous variable under the assumption that @xmath112 is very large , according to standard procedures , we observe that @xmath113 denotes the number of states with energy between @xmath114 and @xmath115 . the density of states is therefore : @xmath116 and the number of states with energy below @xmath11 is @xmath117 and magnetization @xmath118 of a system of @xmath112 non - interacting @xmath100 spins , as predicted by the boltzmann entropy @xmath119 , and the gibbs entropy @xmath120 here @xmath121 . only gibbs magnetization conforms with the physical magnetization @xmath122 . ] figure [ fig : fig1 ] shows the gibbs and boltzmann temperatures and magnetizations as functions of @xmath11 calculated with @xmath123 for larger values of @xmath112 qualitatively similar plots are obtained . a very unphysical property of @xmath124 is that with the flip of a single spin it jumps discontinuously from @xmath125 to @xmath126 in the thermodynamic limit . the usual reply to such a criticism would be , following @xcite , to say that one should look instead at the quantity @xmath127 , which displays no divergence . no way out is however possible if one considers the magnetization . as can be seen from the figure , only @xmath120 reproduces the exact result , eq . ( [ eq : exact ] ) whereas the magnetization given by @xmath119 is drastically off , and even predicts a nonexistent and unphysical phase transition , in the thermodynamic limit , where the magnetization abruptly jumps from @xmath126 to @xmath125 as a single spin flips from @xmath128 to @xmath129 . the results in the figure are also corroborated by analytical calculations . using eqs . ( [ eq : tmb ] ) and ( [ eq : tmg ] ) with eqs . ( [ eq : omega ] ) and ( [ eq : omega ] ) we obtain @xmath130 thus the discrepancy @xmath131 between the boltzmann magnetization and the physical magnetization is given by the negative boltzmann thermal energy rescaled by the applied magnetic field : @xmath132 since @xmath124 diverges around the zero energy in the thermodynamic limit , so does the discrepancy @xmath131 . note that the discrepancy also diverges as the intensity of the applied magnetic field decreases . it is interesting to notice that , while in the thermodynamic limit @xmath133 approaches @xmath124 for @xmath134 , the same is not true for @xmath135 , which distinctly deviates from @xmath136 for both @xmath137 and @xmath134 . this unveils the fact , apparently previously unnoticed , that boltzmann and the gibbs entropy are not equivalent even in the lower part of the spectrum of large spin systems . lines do not coincide with the adiabats @xmath138 . ] equation ( [ eq : mb ] ) is a special case of a general relation linking the boltzmann forces ( @xmath139 ) and the gibbs forces @xmath140 ( i.e. , the thermodynamic forces @xmath141 ) , reading : @xmath142 this equation accompanies a similar relation linking boltzmann and gibbs temperatures @xmath143 with @xmath144 being the heat capacity . equations ( [ eq : mbvsmg ] ) and ( [ eq : tbvstg ] ) follow by taking the derivative with respect to @xmath11 of @xmath22 and @xmath133 respectively . the reason for the thermodynamic inconsistency of @xmath119 ( consistency of @xmath120 ) can also be understood in the following way . consider the heat differential @xmath145 . clearly @xmath146 is an integrating factor : @xmath147 . hence @xmath148 is a primitive . accordingly the adiabats are determined by the equation @xmath149 and the entropy must be some monotonic function of @xmath150 , that is of @xmath151 . by inspecting eqs . ( [ eq : omega ] ) and ( [ eq : omega ] ) we see that the phase volume @xmath2 is a monotonic function of @xmath151 while the density of states @xmath47 is not a function of @xmath151 ; hence @xmath119 is thermodynamically inconsistent . the inequivalence of @xmath120 and @xmath119 is most clearly seen by plotting the iso-@xmath119 lines in the thermodynamic space @xmath152 ; see fig . [ fig : fig2 ] . note that the adiabats , eq . ( [ eq : adiabats ] ) are straight lines passing through the origin . the iso-@xmath119 lines instead predict a completely different structure of the adiabats . note in particular that the iso-@xmath119 lines are closed . this evidences their thermodynamical inconsistency . summing up : the boltzmann entropy severely fails to accomplish one of its basic tasks , namely , reproducing the correct value of the thermodynamic forces and of heat . we have shown that , within the microcanonical formalism there is only one possible choice of entropy that is consistent with the second law and the equation of state of an ideal gas , namely , the gibbs entropy . discarding the gibbs entropy in favour of the boltzmann entropy , may accordingly result in inconsistency with either of those two pillars . for the great majority of large thermodynamic systems , gibbs and boltzmann entropies practically coincide ; hence there is no problem regarding which we choose . however , there are cases when the two do not coincide : examples are spin systems @xcite and point vortex gases @xcite , where boltzmann temperature , in disagreement with gibbs temperature , has no definite sign , and the boltzmann entropy can largely fail to predict correct values of thermodynamic forces . it must be stressed that the demonstrated failure of the boltzmann entropy to reproduce the thermodynamic forces is not restricted to small systems , where the failure was already known to occur @xcite , but survives , and even becomes more prominent , in the thermodynamic limit , where the boltzmann entropy predicts an unphysical and nonexistent phase transition in the magnetization of a system of non - interacting spins in a magnetic field . in the light of the present results , together with the established fact that the gibbs entropy conforms with all thermodynamic laws @xcite , the issue of which entropy expression is correct is apparently now fully and ultimately settled . the author is indebted to jrn dunkel , stefan hilbert , peter talkner and especially peter hnggi , for the many discussions we had on this topic for years . this research was supported by a marie curie intra european fellowship within the 7th european community framework programme through the project nequflux grant no . 623085 and by the cost action no . mp1209 `` thermodynamics in the quantum regime . ''
a question that is currently highly debated is whether the microcanonical entropy should be expressed as the logarithm of the phase volume ( volume entropy , also known as the gibbs entropy ) or as the logarithm of the density of states ( surface entropy , also known as the boltzmann entropy ) . rather than postulating them and investigating the consequence of each definition , as is customary , here we adopt a bottom - up approach and construct the entropy expression within the microcanonical formalism upon two fundamental thermodynamic pillars : ( i ) the second law of thermodynamics as formulated for quasi - static processes : @xmath0 is an exact differential , and ( ii ) the law of ideal gases : @xmath1 . the first pillar implies that entropy must be some function of the phase volume @xmath2 . the second pillar singles out the logarithmic function among all possible functions . hence the construction leads uniquely to the expression @xmath3 , that is the volume entropy . as a consequence any entropy expression other than that of gibbs , e.g. , the boltzmann entropy , can lead to inconsistencies with the two thermodynamic pillars . we illustrate this with the prototypical example of a macroscopic collection of non - interacting spins in a magnetic field , and show that the boltzmann entropy severely fails to predict the magnetization , even in the thermodynamic limit . the uniqueness of the gibbs entropy , as well as the demonstrated potential harm of the boltzmann entropy , provide compelling reasons for discarding the latter at once .
let @xmath1 be a finite type surface , obtained from a closed oriented surface of genus @xmath5 by removing @xmath6 punctures . the ( orientation preserving ) _ mapping class group _ of @xmath1 is denoted @xmath7 where @xmath8 is the group of orientation preserving homeomoprhisms and @xmath9 is the normal subgroup of homeomorphisms isotopic to the identity . the _ teichmller space _ of @xmath1 , denoted @xmath10 , is the space of conformal structures on @xmath1 modulo isotopy , or equivalently the space of complete , finite area hyperbolic structures on @xmath1 modulo isotopy ; the phrase `` modulo isotopy '' means `` modulo the action of @xmath9 '' . thurston s boundary of teichmller space , denoted @xmath11 , is the space of projective whitehead classes of measured foliations on @xmath1 . there is a natural topology on the set @xmath2 which makes it homeomorphic to a ball of dimension @xmath12 with interior @xmath13 and boundary @xmath14 . the group @xmath0 acts naturally on @xmath13 and on @xmath14 , extending to an action by homeomorphisms on @xmath15 . the theme of this survey is that subgroups of @xmath0 can be studied via the geometric / dynamical properties of their action on @xmath2 , just as discrete subgroups of the isometries of hyperbolic space @xmath3 can be studied via their action on @xmath4 . recent problem lists with this theme in mind include @xcite and @xcite . we will survey several topics exemplifying this theme , in the following sections : [ sectionclassification ] : : the thurston bers classification of mapping classes of @xmath0 . [ sectiontits ] : : the tits alternative . [ sectionlimits ] : : limit sets and domains of discontinuity . [ sectionconvexcocompact ] : : convex cocompact subgroups . [ sectionstabilizers ] : : stabilizers of teichmller discs . [ sectioncombinations ] : : combination theorems and surface subgroups . here is a brief description of each of these topics , to be filled out in the rest of the paper . thurston @xcite and bers @xcite classified individual elements of @xmath0 , or equivalently cyclic subgroups of @xmath0 , via the geometry and dynamics of their actions on @xmath15 . finite order mapping classes have fixed points in @xmath13 . infinite order irreducible mapping classes have positive translation distance in @xmath16 , which is minimized along a unique teichmller geodesic , and they act on @xmath17 with source - sink dynamics . finite order mapping classes have fixed points in @xmath16 , at which the translation distance of zero is minimized . and infinite order reducible mapping classes have translation distance in @xmath16 whose infimum is not realized . the _ tits alternative _ , originally formulated and proved for lattices in semisimple lie groups , was also formulated and proved for subgroups of @xmath0 independently by mccarthy @xcite and ivanov @xcite . the tits alternative says that each subgroup of @xmath0 has either an abelian subgroup of finite index , or a free subgroup of rank @xmath18 . _ limit sets and domains of discontinuity _ of subgroups of @xmath0 were studied by mccarthy papadopoulos @xcite , in analogy to limit sets and domains of discontinuity for discrete subgroups of @xmath19 . _ convex cocompact subgroups _ of @xmath0 , including schottky subgroups of @xmath0 , were originally studied by farb and mosher @xcite and further developed by kent leininger @xcite and hamenstadt @xcite , in analogy to convex cocompact subgroups and schottky subgroups of @xmath19 . _ teichmller discs _ in @xmath13 are isometrically embedded copies of the hyperbolic plane , and their stabilizers are analogues in @xmath0 of fuchsian subgroups of @xmath19 , which are stabilizers of isometrically embedded copies of @xmath20 in @xmath3 . the _ leininger reid combination theorem _ @xcite gives a method for building closed surface subgroups of @xmath0 , by combining certain teichmller disc stabilizer subgroups , in analogy to the maskit combination theorem and related results for building discrete subgroups of @xmath19 by combining simpler discrete subgroups . because of our focus on the geometric / dynamic properties of subgroups of @xmath0 , we have nothing to say about the very rich and interesting algebraic tools that are used to study subgroups , particularly the torelli subgroup and the lower central series of @xmath0 . we start by establishing what will be the basic notions and notations for this survey . references for this material include @xcite and @xcite . [ [ teichmllerspace - and - the - thurston - boundary . ] ] teichmller space and the thurston boundary . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + an _ essential curve _ on @xmath1 is a simple closed curve which bounds neither a disc nor a once - punctured disc . the isotopy class of an essential curve @xmath21 is denoted @xmath22 $ ] . let @xmath23 be the set of isotopy classes of essential curves on @xmath1 . an _ essential curve family _ is a finite set of pairwise disjoint essential curves , no two of which are isotopic . the _ teichmller space _ of @xmath1 , denoted @xmath24 , is the set of complete , finite area hyperbolic structures on @xmath1 , or equivalently the set of conformal structures with removable singularities at the punctures , modulo isotopy . isotopy classes , in @xmath23 or in @xmath13 or in other contexts , will be denoted formally with square brackets @xmath25 $ ] , although informally these brackets will often be dropped , as in the next sentence . there is an embedding @xmath26 defined by assigning to each @xmath27 and @xmath28 the minimal length of curves isotopic to @xmath21 in the hyperbolic structure @xmath29 . the topology on @xmath24 is defined so that the map @xmath26 is a homeomorphism onto its image . the group @xmath30 acts naturally on @xmath23 and on @xmath24 , and so the embedding @xmath26 is @xmath0-equivariant . a _ quadratic differential _ on @xmath27 is , formally , a holomorphic section of the symmetric square of the cotangent bundle of @xmath29 , whose integral is finite . such sections can be added and multiplied by complex scalars , and so the quadratic differentials on @xmath29 form a vector space @xmath31 over the complex numbers . informally , to each conformal coordinate @xmath32 for @xmath29 , a quadratic differential @xmath33 assigns an expression @xmath34 , where @xmath35 is holomorphic and such that the expression behaves correctly under coordinate change . for @xmath36 outside of a finite singular set , @xmath33 has a local coordinate @xmath32 in which its expression is @xmath37 and @xmath38 , and @xmath32 is unique up to multiplication by @xmath39 . on each point of the singular set , including each puncture , @xmath33 has a local coordinate @xmath32 in which its expression is @xmath40 for some integer @xmath41 , where @xmath42 if @xmath43 is not a puncture and @xmath44 if @xmath43 is a puncture , and this coordinate @xmath32 is well - defined up to multiplication by a @xmath45 root of unity . the _ horizontal and vertical measured foliations _ of @xmath33 are defined in each regular canonical coordinate @xmath46 as follows : the horizontal foliation has leaves on which @xmath47 is constant , with transverse measure @xmath48 ; the vertical foliation has leaves on which @xmath49 is constant , with transverse measure @xmath50 . at a singularity @xmath43 near which @xmath33 has the canonical expression @xmath40 we say that the horizontal and vertical measured foliations have _ @xmath41-pronged singularities _ at @xmath43 . teichmller s theorem says that for any @xmath51 there exist unique choice of quadratic differentials @xmath33 on @xmath29 called the _ initial quadratic differential _ , @xmath52 on @xmath53 called the _ terminal quadratic differential _ , and a number @xmath54 , such that if @xmath46 is a regular local coordinate for @xmath33 then ( up to isotopy of @xmath53 ) the expression @xmath55 defines a regular local coordinate for @xmath52 . this number @xmath56 is called the _ teichmller distance _ between @xmath29 and @xmath53 , and it defines a metric on @xmath24 which gives the same topology as that defined above . composing the embedding @xmath26 with the projectivization @xmath57 , the composition @xmath58 is still an embedding , and its image is precompact . the closure of the image minus the image defines _ thurston s boundary _ for @xmath24 , and this boundary is characterized as follows . a _ measured foliation _ on @xmath1 is a foliation with a finite singular set , equipped with a positive transverse borel measure , so that at each singularity @xmath43 including punctures there exists @xmath44 such that the singularity is locally modelled on a @xmath41-pronged singularity of @xmath40 , where @xmath42 if @xmath43 is not a puncture . there is an equivalence relation on measured foliations generated by isotopy and by the _ whitehead move _ , a move which collapses to a point each leaf segment whose endpoints are a pair of singularities at least one of which is not a puncture . the equivalence classes are called _ whitehead classes _ , and the set of whitehead classes of measured foliations is denoted @xmath59 ( again we use @xmath25 $ ] to denote whitehead class , except that the notation is often dropped ) . there is an embedding @xmath60 defined by assigning to each @xmath61 and @xmath28 the number @xmath62 which is the minimal integral of curves homotopic to @xmath21 with respect to the transverse measure on @xmath63 . the topology on @xmath59 is chosen so that the map @xmath60 is a homeomorphism onto its image . when the transverse measure on @xmath63 is multiplied by a number @xmath64 we denote the result @xmath65 , and this generates the relation of _ projective equivalence _ on @xmath59 . the space of projective equivalence classes is denoted @xmath14 , and the resulting map @xmath66 is an embedding . thurston s compactification theorem says that the disjoint union of @xmath13 and @xmath14 embeds into @xmath67 , the image is a closed ball of dimension @xmath68 , the interior of this ball is the image of @xmath13 , and its boundary sphere is the image of @xmath14 . there is a natural action of @xmath0 on @xmath14 with respect to which the embedding @xmath69 is equivariant , and hence @xmath0 acts homeomorphically on the ball @xmath2 . consider an essential closed curve @xmath28 . the _ enlargement _ of @xmath21 is a measured foliation @xmath70 well - defined up to whitehead equivalence as follows . first one chooses a spine for @xmath71 containing all punctures . the complement of this spine is an open annulus with core @xmath21 . then one foliates this annulus by curves isotopic to @xmath21 , which together with the spine defines @xmath70 as a singular foliation . the transverse measure on @xmath70 is defined so that a curve in the annulus from boundary to boundary transverse to @xmath70 has total transverse measure 1 . this construction generalizes to the _ enlargement of a weighted multicurve _ , a formal sum @xmath72 where @xmath73 is an essential curve system and @xmath74 ; in this case one chooses a spine for @xmath75 containing all punctures , whose complement is a union of open annuli with cores @xmath76 respectively ; then one foliates these annuli by closed curves parallel to the cores ; then one assigns total transverse measures @xmath77 , respectively , to these foliated annuli . the set of points in @xmath14 represented by enlargements of weighted multicurves is dense . in fact , the set of points in @xmath59 represented by enlargements of essential closed curves is dense . the intersection pairing @xmath78 taking @xmath79 to @xmath80 extends to a continuous intersection pairing @xmath81 , called the _ intersection number _ of @xmath82 ; uniqueness of the extension follows from denseness of @xmath23 in @xmath59 . a pair of measured foliations @xmath82 is said to _ jointly fill _ the surface @xmath1 if for every @xmath28 either @xmath80 or @xmath83 is nonzero , and @xmath84 is also nonzero . this happens if and only if there exists a conformal structure @xmath27 and a quadratic differential @xmath33 on @xmath29 such that @xmath82 are whitehead equivalent to the horizontal and vertical measured foliations of @xmath33 ; moreover , @xmath29 and @xmath33 are uniquely determined by @xmath82 . the set of geodesic lines in @xmath13 can be parameterized as follows . choose measured foliations @xmath82 which jointly fill . for each @xmath85 , the measured foliations @xmath86 jointly fill and determined a conformal structure @xmath87 . the map @xmath88 is a geodesic embedding @xmath89 . the image of this geodesic depends only on the projective classes @xmath90 $ ] , @xmath91 $ ] , and we say that this is the _ teichmller geodesic _ with directions @xmath90 $ ] , @xmath91 $ ] . the set of geodesic rays in the teichmller metric can be parameterized as follows . first we use a theorem of hubbard and masur @xcite which says that for each @xmath27 and @xmath90 \in { \mathcal{pmf}}$ ] there is up to real scalar multiple a unique @xmath92 whose horizontal measured foliation is in the given projective class @xmath90 $ ] . for each @xmath93 we can then transform each regular canonical coordinate @xmath46 of @xmath33 to get a coordinate @xmath55 defining an isometric embedding @xmath94 . the image of this map is called the _ teichmller ray with basepoint @xmath29 and direction @xmath90$]_. a measured foliation @xmath63 is _ arational _ if @xmath95 for every @xmath28 . equivalently , there is no closed loop in @xmath1 which is everywhere parallel to @xmath63 , nor is there any closed path from puncture to puncture which is everywhere parallel to @xmath63 . if @xmath82 are arational , and if @xmath96 , then @xmath63 , @xmath97 jointly fill @xmath1 . [ [ virtual - torsion - free . ] ] virtual torsion free . + + + + + + + + + + + + + + + + + + + + + serre @xcite originally proved that the group @xmath0 has a finite index subgroup which is torsion free . for another proof see @xcite corollary 1.5 . [ theoremvtf ] the kernel of the homomorphism @xmath98 is torsion free . this kernel has index bounded above by @xmath99 which is finite because @xmath100 has finite rank @xmath101 over @xmath102 . note that the bound in the index of the kernel is exponential in @xmath103 and @xmath43 . [ [ gromov - hyperbolic - metric - spaces . ] ] gromov hyperbolic metric spaces . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + much of this survey is motivated by analogues in the theory of discrete groups of isometries of gromov hyperbolic metric spaces , particularly hyperbolic spaces @xmath3 . we recall some of the notions of this theory , whose details can be found in , for example , @xcite , @xcite , @xcite , @xcite . let @xmath104 be a metric space . @xmath104 is _ proper _ if closed balls are compact . @xmath104 is a _ geodesic metric space _ if for any @xmath105 there exists an isometrically embedded path @xmath106 \to x$ ] , a _ geodesic _ , from @xmath49 to @xmath47 . we ll use the notation @xmath107 $ ] to refer to the image of some geodesic . the _ hausdorff distance _ between two sets @xmath108 is the infimum of @xmath109 such that @xmath110 and @xmath111 . a geodesic metric space @xmath104 is _ hyperbolic _ in the sense of gromov if there exists @xmath109 such that for any @xmath112 and any geodesics @xmath107 $ ] , @xmath113 $ ] , @xmath114 $ ] in @xmath104 , @xmath107 $ ] is contained in the @xmath115-neighborhood of @xmath113 { \cup}[z , x]$ ] . the _ boundary _ @xmath116 is defined to be the set of geodesic rays modulo the relation of finite hausdorff distance ; given a geodesic ray @xmath117 , let @xmath118 $ ] denote the corresponding point of @xmath116 . there is a natural _ gromov topology _ on @xmath119 , whose definition we shall not recount , but we shall mention several important properties : * if @xmath104 is a proper geodesic metric space then @xmath116 and @xmath119 are compact . * every geodesic ray @xmath117 converges in @xmath119 to @xmath118 $ ] . * if @xmath104 is proper then for any @xmath120 and any @xmath121 there exists a geodesic ray @xmath117 based at @xmath49 and converging to @xmath122 , and for any @xmath123 there exists a bi - infinite geodesic whose two ends converge to @xmath124 . of course , the primary example of the gromov topology is the union of @xmath3 with its sphere at infinity @xmath125 . when @xmath104 is not proper then the compactness properties above can fail . another important example for us of a gromov hyperbolic metric space is the curve complex of @xmath1 , discussed below . the discrete cyclic groups of isometries of @xmath3 or of any proper , geodesic , gromov hyperbolic metric space @xmath104 fall into three classes : _ elliptic _ groups which are finite and which fix a point of @xmath104 ; _ parabolic _ groups which are infinite cyclic and which fix a point of @xmath116 but no point of @xmath104 ; and _ loxodromic _ groups which are infinite cyclic and which translate along a bi - infinite geodesic of @xmath104 . thurston discovered an analogous classification of elements of @xmath0 @xcite ; see also @xcite and @xcite . work of bers strengthens the analogy in terms of the action on @xmath13 @xcite . there are two major classifications of a mapping class @xmath126 : whether @xmath127 is finite order ; and whether @xmath127 is _ reducible _ meaning that there exists an essential curve family @xmath23 , called a _ reduction family _ for @xmath127 , such that @xmath128=[{\mathcal c}]$ ] . if @xmath127 is not reducible then it is _ irreducible_. a homeomorphism @xmath129 is said to be _ pseudo - anosov _ if there exists a transverse pair of measured foliations @xmath130 , @xmath131 which jointly fill the surface , called the _ stable and unstable measured foliations _ of @xmath132 , and there exists a number @xmath133 called the _ expansion factor _ of @xmath132 , such that @xmath132 takes each leaf of @xmath130 to a leaf of @xmath130 , each leaf of @xmath131 to a leaf of @xmath131 , @xmath132 contracts the leaves of @xmath130 by a factor of @xmath134 with respect to the @xmath131 measure , and @xmath132 expands the leaves of @xmath131 by a factor of @xmath134 with respect to the @xmath130 measure . to say this more briefly , @xmath135 and @xmath136 . a mapping class @xmath137 is said to be pseudo - anosov if it is represented by a pseudo - anosov homeomorphism . thurston proved that every mapping class @xmath137 falls into one of three types : finite order , reducible , or pseudo - anosov ; in other words , every infinite order , irreducible mapping class is _ pseudo - anosov_. the first two types are not mutually exclusive , but a mapping class is pseudo - anosov if and only if it is infinite order and irreducible . there are several necessary and/or sufficient conditions for the three types of mapping classes , expressed in terms of the geometry and dynamics of the action on compactified teichmller space . [ [ finite - order - mapping - classes . ] ] finite order mapping classes . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a mapping class @xmath137 is of finite order if and only if it has a fixed point in @xmath13 , in which case there exists a conformal structure @xmath29 on @xmath1 and a homeomorphism @xmath132 representing @xmath127 such that @xmath138 . it follows that @xmath132 is a finite order homeomorphism . a sufficient condition for @xmath137 to be of finite order , proved in @xcite , is the existence of an arational measured foliation @xmath63 such that @xmath139=[{{\mathcal f}}]$ ] . it is possible for a finite order mapping class @xmath127 to be irreducible . this happens if and only if , for @xmath132 a finite order homeomorphism representing @xmath127 , the quotient orbifold @xmath140 is a sphere with three cone points . there is a bound on the order of a finite order element of @xmath30 , in fact a bound on the order of a finite subgroup @xmath141 , depending only on the topology of @xmath1 . one way to see this is by applying theorem [ theoremvtf ] , which gives a bound which is exponential in @xmath103 and @xmath43 . a linear bound is obtained using the much deeper theorem of gabai @xcite and casson jungreis @xcite which says that @xmath142 is realized as a finite group of homeomorphisms of @xmath1 , also denoted @xmath142 . we therefore have @xmath143 , where the denominator is the orbifold euler characteristic of the quotient orbifold @xmath144 . the numerator and denominator of this fraction being negative , @xmath145 is maximized when @xmath146 is maximized . the maximum value of the euler characteristic of a compact , oriented hyperbolic 2-orbifold is realized by the spherical orbifold with three cone points of angles @xmath147 , whose euler characteristic equals @xmath148 . it follows that @xmath149 when @xmath1 is compact . when @xmath1 has one or more punctures there is a slightly better bound , because the maximum value of the euler characteristic of a noncompact , finite area , oriented hyperbolic 2-orbifold is acheived by the modular space of @xmath150 , a once - punctured sphere with two cone points of angles @xmath151 and with euler characteristic @xmath152 , and so @xmath153 . to summarize : [ theoremfinitebound ] if @xmath142 is a finite subgroup of @xmath0 then @xmath154 [ [ pseudo - anosov - mapping - classes . ] ] pseudo - anosov mapping classes . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + bers proved @xcite that a mapping class @xmath137 is pseudo - anosov if and only if the infimum of @xmath155,\phi[\sigma])$ ] is positive and is acheived by some @xmath156 \in { { { \cal t}}}$ ] . in this case , by combining with the fact that the stable and unstable foliations @xmath157 are uniquely ergodic , it follows that the set of points @xmath156 $ ] where the infimum is acheived is a bi - infinite geodesic in @xmath13 along which @xmath127 translates , with translation distance equal to @xmath158 . this geodesic is called the _ axis _ for @xmath127 in @xmath13 . this axis has two ideal endpoints in @xmath14 , namely @xmath159 $ ] and @xmath160 $ ] . the action of @xmath127 on @xmath15 is a _ source sink action _ , with _ source _ @xmath159 $ ] and _ sink _ @xmath159 $ ] , meaning that for every point @xmath161,p[{{\mathcal f}}^u]\}$ ] , @xmath162 \quad\text{and}\quad \lim_{i \to + { \infty } } \phi^i(x ) = p[{{\mathcal f}}^u]\ ] ] it follows that the fixed point set of @xmath127 in @xmath15 is @xmath163,p[{{\mathcal f}}^u]\}\ ] ] a sufficient ( and necessary ) condition for @xmath137 to be pseudo - anosov is the existence of an arational measured foliation @xmath63 and @xmath164 , @xmath165 such that @xmath139=\lambda[{{\mathcal f}}]$ ] ; if this happens then @xmath90 = p[{{\mathcal f}}^s]$ ] or @xmath160 $ ] . see @xcite for the proof . [ [ reducible - mapping - classes . ] ] reducible mapping classes . + + + + + + + + + + + + + + + + + + + + + + + + + + if a mapping class @xmath127 is reducible , and if @xmath23 is a reducing family for @xmath127 , then there exists a representative @xmath132 of @xmath127 such that @xmath166 . given a component @xmath167 of @xmath168 , let @xmath169 be the first return time of @xmath132 to @xmath167 , meaning that @xmath170 and @xmath171 for @xmath172 . then the first return time @xmath173 and the mapping class of the first return map @xmath174 are determined by @xmath127 , independent of the representative @xmath132 . these mapping classes are called the _ component mapping classes _ of @xmath127 relative to @xmath23 . a reducing family @xmath23 for a mapping class @xmath127 is _ complete _ if all of the component mapping classes of @xmath127 relative to @xmath23 are finite order or pseudo - anosov . a minimal , complete reducing family @xmath23 for @xmath127 is unique up to isotopy ; see for example @xcite , @xcite . if @xmath23 , @xmath175 are any two minimal , complete reducing families for @xmath127 , if @xmath176 are representatives of @xmath127 such that @xmath166 , @xmath177 , and if @xmath178 satisfies @xmath179 , then @xmath180 conjugates the component mapping classes of @xmath127 relative to @xmath23 to the component mapping classes of @xmath127 relative to @xmath175 . we may therefore speak , in a well - defined manner , of the _ component mapping classes of @xmath127 _ , meaning the component mapping classes of @xmath127 relative to any minimal , complete reducing family . bers proved @xcite that a mapping class @xmath127 is infinite order and reducible if and only if the infimum of @xmath155,\phi[\sigma])$ ] is not acheived by any @xmath156 \in { { { \cal t}}}$ ] . moreover , two subcases are distinguished : if the infimum equals zero then each component mapping class of @xmath127 has finite order ; whereas if the infimum is positive then some component mapping class of @xmath127 is pseudo - anosov . a sufficient ( and necessary ) condition for @xmath137 to be reducible is the existence of a non - arational measured foliation @xmath63 such that @xmath139={{\mathcal f}}$ ] . a complete description of the dynamics of a reducible mapping class acting on @xmath181 is obtained in @xcite . [ [ independence - of - pseudo - anosov - homeomorphisms . ] ] independence of pseudo - anosov homeomorphisms . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in a discrete group acting on @xmath3 , given two loxodromic elements @xmath182 , either @xmath183 or @xmath184 . the same is true for pseudo - anosov homeomorphisms . proofs are given in @xcite . we shall give a proof here . [ lemmadisjointfixed ] for any pseudo - anosov @xmath185 , either @xmath183 or @xmath184 . in the latter case we say that @xmath182 are _ independent_. supposing that @xmath186 , up to inverses of @xmath182 and up to isotopy we may assume that @xmath187 , and that @xmath139 = \lambda[{{\mathcal f}}]$ ] and @xmath188=\lambda'[{{\mathcal f}}]$ ] for some @xmath189 . let @xmath190 be the intersection of @xmath191)$ ] and the kernel of the homomorphism @xmath98 , and so by theorem [ theoremvtf ] it follows that @xmath190 has finite index in @xmath191)$ ] . it suffices to prove that @xmath190 is infinite cyclic , for in that case @xmath182 have positive powers in @xmath190 , implying that @xmath192 for some @xmath193 , implying that @xmath194 = [ { { \mathcal f}}^s(\phi^m ) ] = [ { { \mathcal f}}^s(\phi'{}^n ) ] = [ { { \mathcal f}}^s(\phi')]$ ] . consider the homomorphism @xmath195 which is characterized by the equation @xmath196 = \ell(\psi)[{{\mathcal f}}]$ ] . suppose that @xmath197 . as seen above , @xmath198 has finite order . but @xmath199 and so by theorem [ theoremvtf ] , @xmath198 is the identity . the homomorphism @xmath200 is thus injective . but the image of @xmath200 is discrete indeed , the set of expansion factors of pseudo - anosov elements of @xmath30 is bounded away from @xmath201 , because each is an eigenvalue of a square integer matrix of bounded size @xcite . it follows that @xmath190 is infinite cyclic . [ [ the - ping - pong - argument - and - free - subgroups . ] ] the ping - pong argument and free subgroups . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + here is the central construction of the tits alternative : [ lemmapsanpingpong ] for any two independent pseudo - anosov mapping classes @xmath202 , for sufficiently large integers @xmath203 the elements @xmath204 freely generate a free subgroup of @xmath30 . the proof of this lemma is a standard ping - pong argument , using that @xmath202 each act with source - sink dynamics on @xmath13 , and that their fixed points sets are disjoint . a general ping - pong lemma stated in a topological context is as follows : let @xmath142 act on a topological space @xmath104 , let @xmath205 , and suppose that there exists a pairwise disjoint collection of open sets @xmath206 , @xmath207 such that for each @xmath208 and each @xmath209 we have @xmath210 then @xmath142 is freely generated by @xmath211 . the way that lemma [ lemmapsanpingpong ] is proved is to note that when @xmath202 are independent pseudo - anosov mapping classes , and if @xmath212 , @xmath213 , @xmath214 , @xmath215 are pairwise disjoint neighborhoods of the fixed points @xmath216 $ ] , @xmath217 $ ] , @xmath218 $ ] , @xmath219 $ ] , then by applying the source sink dynamics of @xmath127 and of @xmath198 , the hypotheses of the ping pong lemma are satisfied for @xmath204 as long as @xmath220 are sufficiently large . one of the main themes of the rest of this survey is that by taking still more care with the proof of lemma [ lemmapsanpingpong ] , one can show that the free subgroup @xmath221 satisfies stronger geometric properties reminiscent of convex cocompact discrete subgroups of hyperbolic isometries . [ [ the - virtual - centralizer - and - virtual - normalizer - of - a - pseudo - anosov - cyclic - subgroup . ] ] the virtual centralizer and virtual normalizer of a pseudo - anosov cyclic subgroup . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + consider a group @xmath142 and a subgroup @xmath222 . the _ virtual centralizer _ @xmath223 of @xmath224 in @xmath142 is the subgroup of all @xmath225 which commute with a finite index subgroup of @xmath224 . the _ virtual normalizer _ @xmath226 is the subgroup of all @xmath225 such that @xmath227 has finite index in @xmath228 and in @xmath224 . the virtual centralizer and virtual normalizer of an infinite cyclic pseudo - anosov subgroup each have a particularly nice geometric / dynamic description , and the proof of this description uses the ping - pong argument in a nice way . given an action of a group @xmath142 on a space @xmath104 , for a subspace @xmath229 denote @xmath230 . if @xmath1 is homeomorphic to a line denote @xmath231 to be the subgroup of @xmath232 that preserves the orientation on @xmath1 . [ theoremvcvn ] given a pseudo - anosov @xmath137 with stable and unstable foliations @xmath233 $ ] , @xmath234 $ ] and teichmller axis @xmath235 , we have : @xmath236 ) = \operatorname{stab}(p[{{\mathcal f}}^u ] ) = \operatorname{stab}_+(a ) \\ vn{\langle}\phi{\rangle}&= \operatorname{stab}\{p[{{\mathcal f}}^s],p[{{\mathcal f}}^u]\ } = \operatorname{stab}(a)\end{aligned}\ ] ] the group @xmath237 has index at most @xmath238 in @xmath239 , it has an infinite cyclic subgroup of finite index , and the index in @xmath237 of the maximal such subgroup is bounded by the maximum order of a finite subgroup of @xmath30 . the group @xmath239 acts properly discontinuously on the teichmller geodesic @xmath235 and so is virtually cyclic . the inclusions @xmath240 and @xmath241 follow immediately . the inclusions @xmath242 ) , \operatorname{stab}(p[{{\mathcal f}}^u ] ) \subset \operatorname{stab}_+(a)$ ] and the equation @xmath243,p[{{\mathcal f}}^u]\ } = \operatorname{stab}(a)$ ] are obvious . suppose that @xmath244,p[{{\mathcal f}}^u]\}$ ] . then @xmath245 is pseudo - anosov , and it is independent of @xmath127 by lemma [ lemmadisjointfixed ] . some powers of @xmath127 and @xmath246 therefore generate a free group of rank 2 , and hence @xmath247 . suppose next that @xmath248)$ ] , and again consider the pseudo - anosov mapping class @xmath245 . if @xmath249 ) \ne p[{{\mathcal f}}^s]$ ] then @xmath246 independent of @xmath127 and @xmath247 , so @xmath250 . if @xmath249 ) = p[{{\mathcal f}}^s]$ ] but @xmath251 ) \ne p[{{\mathcal f}}^u]$ ] then @xmath252 , @xmath253 are not disjoint and not equal , violating lemma [ lemmadisjointfixed ] . and if @xmath198 interchanges @xmath160 $ ] and @xmath159 $ ] then @xmath246 translates along @xmath235 in the direction opposite to @xmath127 and so @xmath254 . the following theorem was proved independently by ivanov @xcite and mccarthy @xcite : for each finite type surface @xmath1 there exist constants @xmath255 such that for any subgroup @xmath142 of @xmath30 , either @xmath142 contains a free subgroup of rank @xmath238 , or @xmath142 contains an abelian subgroup of index @xmath256 and rank @xmath257 . ivanov gave an underlying geometric description of each subgroup of @xmath30 , from which the above tits alternative follows quickly . a subgroup @xmath141 is _ reducible _ if there exists an essential curve family which is invariant under the action of each element of @xmath142 . if no such curve family exists then @xmath142 is _ irreducible_. [ theoremivanov ] if @xmath142 is an infinite , irreducible subgroup of @xmath30 then @xmath142 contains a pseudo - anosov element . for the proof see ivanov s book @xcite . let @xmath142 be an arbitrary subgroup of @xmath0 . suppose first that @xmath142 is irreducible . applying theorem [ theoremivanov ] , @xmath142 contains a pseudo - anosov mapping class @xmath127 . if @xmath258 then by theorem [ theoremvcvn ] the group @xmath142 has an infinite cyclic subgroup of index at most twice the maximum size of a finite subgroup of @xmath0 . if @xmath259 , say @xmath260 , then by lemma [ lemmadisjointfixed ] @xmath261 is a pseudo - anosov element independent of @xmath127 and so by the ping pong argument @xmath142 contains a free group of rank 2 . suppose now that @xmath142 is reducible . let @xmath23 be a maximal essential curve family invariant under @xmath142 . the group @xmath142 acts on the set of oriented elements of @xmath23 ; let @xmath262 denote the kernel of this action , a finite index subgroup of @xmath142 whose index is bounded above by @xmath263 . letting @xmath264 be the components of @xmath168 , it follows that @xmath262 preserves each @xmath265 up to isotopy , and so there is a well - defined restriction homomorphism @xmath266 whose image we denote @xmath267 . by maximality of @xmath23 it follows that @xmath267 is irreducible . if @xmath267 has two independent pseudo - anosov elements then @xmath267 has a free subgroup of rank @xmath18 , and so therefore does @xmath142 . we have reduced to the case that @xmath267 is either finite or is contained in @xmath268 for some pseudo - anosov element @xmath269 . in this case we find a free abelian subgroup of finite index in @xmath142 . let @xmath270 denote either the trivial subgroup or an infinite cyclic group , with index bounded as in theorem [ theoremfinitebound ] or [ theoremvcvn ] . consider the subgroup @xmath271 . the number of subsurfaces @xmath265 is bounded in terms of @xmath103 and @xmath43 , and so we obtain a finite bound on the index of @xmath224 in @xmath142 , in terms of @xmath103 and @xmath43 . restricted to each @xmath265 , the subgroup @xmath224 is either trivial or infinite cyclic , in the latter case generated by a pseudo - anosov element of @xmath265 . @xmath224 also contains dehn twists about the curves in @xmath23 . these pseudo - anosov subsurface mapping classes and the dehn twists all commute with each other . thus @xmath224 is a free abelian group , whose rank is bounded by @xmath272 , which is bounded in terms of @xmath103 and @xmath43 . note that this proof gives a complete description of all free abelian subgroups of @xmath30 : each irreducible one is infinite cyclic generated by a pseudo - anosov mapping class ; and each reducible one is freely generated by dehn twists along curves @xmath23 and pseudo - anosov mapping classes on components @xmath264 of @xmath168 . to compute the maximal rank exactly , note that on any subsurface @xmath265 that supports a pseudo - anosov homeomorphism , one can replace that homeomorphism by a dehn twist , enlarging @xmath23 without decreasing the rank . the maximum rank is thus acheived when each @xmath265 is a pair of pants , in which case the rank equals the maximum value of @xmath273 , which equals @xmath274 . given a gromov hyperbolic , geodesic metric space @xmath104 and a finitely generated , discrete subgroup @xmath275 , its _ limit set _ @xmath276 is , by definition , the accumulation set in @xmath116 of any orbit of @xmath142 in @xmath104 . as long as @xmath142 is nonelementary ( that is , not virtually abelian ) , the limit set may be characterized as the closure of the set of fixed points of loxodromic elements of @xmath142 , or as the unique minimal closed nonempty @xmath142-invariant subset of @xmath116 . assuming in addition that @xmath104 is proper , it follows that @xmath277 is compact , and the _ domain of discontinuity _ @xmath278 is characterized as the unique maximal open @xmath142-invariant set on which the action of @xmath142 is properly discontinuous . in @xcite , masur investigated notions of the limit set and domain of discontinuity for the _ handlebody subgroup _ of @xmath30 when @xmath1 is the boundary of a compact genus @xmath103 handlebody @xmath224 ; this subgroup consists of the mapping classes of homeomorphisms of @xmath1 that extend to @xmath224 . in @xcite , mccarthy and papadopoulos generalized masur s results to arbitrary subgroups of @xmath30 . the results are not quite as pretty as they are for discrete groups of hyperbolic isometries . but as we shall see , the results of kent and leininger @xcite show that limit sets of convex cocompact subgroups of @xmath0 behavior very much like their counterparts in hyperbolic geometry . we start with some elementary cases . when @xmath142 is a finite subgroup of @xmath0 , the natural choice of the limit set is @xmath279 , and the action of @xmath142 on @xmath280 is properly discontinuous . when @xmath142 is virtually cyclic , containing a pseudo - anosov cyclic subgroup @xmath281 with finite index , let @xmath282,p[{{\mathcal f}}^u_\phi]\}$ ] . in the case where the action of @xmath142 does not fix each element of @xmath277 , that is when @xmath283 is an index 2 subgroup of @xmath284 , then @xmath277 is the unique nonempty minimal closed @xmath142-invariant subset of @xmath14 . in the other case , when @xmath285 , then the two points of @xmath277 are the only nonempty minimal closed @xmath142-invariant subsets of @xmath14 . in either case , the action of @xmath142 on @xmath286 is properly discontinuous . consider now the case where @xmath142 is an infinite , irreducible subgroup of @xmath0 and @xmath142 has no finite index cyclic subgroup . define @xmath287 to be the set of fixed points of pseudo - anosov elements of @xmath142 . define @xmath288 to be the closure of @xmath289 . define @xmath290 to be the `` zero set '' of @xmath277 , the set of all @xmath90 \in { \mathcal{pmf}}$ ] such that for some @xmath91 \in \lambda_g$ ] we have @xmath291,[{{\mathcal f}}']{\rangle}= 0 $ ] ; this is well - defined independent of the choice of @xmath292,[{{\mathcal f}}']$ ] in their projective whitehead classes . define @xmath293 . [ theoremmcpapairred ] for @xmath142 an infinite , irreducible subgroup of @xmath0 , the set @xmath277 is the unique minimal nonempty closed @xmath142-invariant subset of @xmath14 . if @xmath294 then @xmath295 , and the action of @xmath142 on @xmath296 is properly discontinuous . we turn now to the case that @xmath142 is an infinite , reducible subgroup of @xmath30 . in general there will be no unique closed minimal @xmath142-invariant subset of @xmath14 . for example , if @xmath142 is a free abelian group generated by dehn twists about the curves in an essential curve family @xmath23 , then @xmath142 has many , many fixed points , namely any point in @xmath14 having zero intersection number with each curve in @xmath23 . nevertheless , as mccarthy and papadopoulos show , there is still a reasonable candidate for a limit set and a domain of discontinuity . let @xmath23 be an essential curve family invariant under @xmath142 such that if @xmath264 is the set of components of @xmath168 , then the restriction @xmath297 is an irreducible subgroup of @xmath298 . as is shown in , say , @xcite , there exists a unique such family @xmath23 which is minimal with respect to inclusion . reindex the @xmath265 so that @xmath299 are the surfaces on which the restriction groups @xmath267 are infinite . each of the groups @xmath267 for @xmath300 therefore contains a pseudo - anosov element of @xmath298 . the stable and unstable foliations of pseudo - anosov elements of @xmath267 , being measured foliations on @xmath265 , may by enlargement be regarded as measured foliations on @xmath1 ; let @xmath301 be the set of all points in @xmath14 obtained in this manner . let @xmath302 be the closure of @xmath303 . define @xmath304 \right)\ ] ] define the zero set @xmath290 and the domain @xmath305 as above . for each infinite , reducible subgroup of @xmath0 , the set @xmath277 is a nonempty closed @xmath142-invariant subset of @xmath14 . if @xmath294 then @xmath295 , and the action of @xmath142 on @xmath296 is properly discontinuous . the set @xmath277 shall be called the _ limit set _ and @xmath296 the _ domain of discontinuity _ of @xmath142 . given a finitely generated , discrete subgroup @xmath306 , the _ convex hull _ of its limit set @xmath277 is the smallest closed , convex subset @xmath307 whose set of accumulation points in @xmath125 is @xmath277 . the action of @xmath142 on @xmath3 preserves @xmath308 , and @xmath142 is said to be _ convex cocompact _ if @xmath142 acts cocompactly on @xmath308 . several conditions are equivalent to convex cocompactness : some ( every ) orbit of @xmath142 in @xmath308 is quasiconvex ; the action of @xmath142 on @xmath296 is cocompact ; the action of @xmath142 on @xmath309 is cocompact . moreover , if @xmath142 is convex cocompact then @xmath142 is word hyperbolic , there is a @xmath142-equivariant homeomorphism from the gromov boundary @xmath310 to @xmath277 , and this homeomorphism extends to a @xmath142-equivariant continuous map from the gromov compactification @xmath311 to @xmath312 . these notions can be generalized with care to a finitely generated , discrete group @xmath142 of isometries of any gromov hyperbolic , geodesic metric space @xmath104 . the theory of convex cocompact subgroups of @xmath0 , originated in @xcite and further developed in @xcite , and @xcite , gives analogues to these results about convex cocompact subgroups of @xmath0 . in addition , the theory develops further equivalent characterizations of convex compactness that are special to the setting of @xmath0 and are important for applications . thus far , the only examples of convex cocompact subgroups of @xmath0 are free @xcite or virtually free examples as constructed by honglin min in her dissertation @xcite . we define a subset @xmath313 to be _ quasiconvex _ if there exist a constant @xmath314 such that for every @xmath105 , each point of the teichmller geodesic @xmath107 $ ] is within distance @xmath314 of some point of @xmath104 . given a closed subset @xmath315 , if every pair @xmath316 jointly fills @xmath1 then we define the _ weak hull _ of @xmath317 to be the subset of @xmath13 defined by @xmath318 and we say that the _ weak hull of @xmath317 is defined_. otherwise , if there exist @xmath316 which do not jointly fill @xmath1 then we say that the weak hull of @xmath317 is undefined . for any finitely generated subgroup @xmath319 , the following are equivalent . * every ( some ) orbit of @xmath142 on @xmath0 is quasiconvex in @xmath13 . * @xmath142 is word hyperbolic , and there exists a @xmath142-equivariant embedding @xmath320 with image denoted @xmath317 , such that the weak hull of @xmath317 is defined , the action of @xmath142 on @xmath321 is cocompact , and the extension of the map @xmath320 by any @xmath142-equivariant map @xmath322 is a continuous map @xmath323 . a group satisfying these conditions is called _ convex cocompact_. when @xmath142 is convex cocompact then every infinite order element of @xmath142 is pseudo - anosov . the connection of convex cocompactness with the limit set @xmath277 and domain of discontinuity @xmath296 as defined by mccarthy - papadapoulos is given in the following results of kent and leininger : for any finitely generated subgroup @xmath319 , the following are equivalent : * @xmath142 is convex cocompact . * the weak hull @xmath324 of the mccarthy - papadapoulos limit set @xmath277 is defined , and the action of @xmath142 on @xmath324 is cocompact . * the action of @xmath142 on the mccarthy - papadapoulos domain of discontinuity @xmath296 is cocompact . * the action of @xmath142 on @xmath325 is cocompact . furthermore , if @xmath142 is convex cocompact , then the image of the embedding @xmath326 is the mccarthy - papadapoulos limit set @xmath277 , and @xmath327 , and so @xmath328 . [ [ convex - cocompactness - on - the - curve - complex . ] ] convex cocompactness on the curve complex . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the _ curve complex _ @xmath329 is the simplicial complex whose vertex set is @xmath23 , and whose @xmath173-simplices are the isotopy classes of essential curve families of cardinality @xmath330 . to be precise , a set of distinct vertices @xmath331,\ldots,[c_n ] \in { \mathcal c}$ ] spans an @xmath173-simplex of @xmath332 if and only if there are representatives @xmath333 which are pairwise disjoint . the complex @xmath332 is locally infinite , in fact the link of every simplex of positive codimension is infinite . the natural action of @xmath0 on @xmath332 has finitely many orbits of simplices . the number of orbits of vertices is @xmath334 when @xmath335 there is one orbit of nonseparating curves , but in genus zero every curve is separating . the orbit of a separating curve is determined by how it partitions the genus and the number of punctures . when @xmath336 , the relevant partition is @xmath337 with @xmath338 , and when @xmath339 the relevant partition @xmath340 with @xmath341 . when @xmath335 , @xmath342 then the relevant partition is an unordered partition of the ordered pair @xmath343 , where the four values @xmath344 , @xmath345 , @xmath346 , @xmath347 are forbidden but all other values with @xmath348 and @xmath349 are allowed . the large scale geometric properties of @xmath332 are of @xmath0 are related in the following manner . given a finitely generated group @xmath142 and a finite collection of finitely generated subgroups @xmath350 , assume that the generators of @xmath142 include generators of each of the subgroups @xmath350 , so that the cayley graph of @xmath351 embeds , @xmath351-equivariantly , in the cayley graph @xmath190 of @xmath142 . define the _ coned off cayley graph _ of @xmath142 relative to @xmath350 by adding a new vertex to @xmath190 for each left coset @xmath352 of each @xmath351 , and adding new edges connecting the new vertex to all the vertices of @xmath352 . we say that @xmath190 is _ weakly hyperbolic _ relative to @xmath350 if the coned off cayley graph is a gromov hyperbolic metric space . choose unique representatives @xmath353,\ldots,[c_m]$ ] for the orbits of the action of @xmath0 on @xmath23 , and let @xmath354 { < } { { \mathcal m \mathcal c \mathcal g}}$ ] . it is easy to see @xcite that @xmath332 is equivariantly quasi - isometric to the coned off cayley graph of @xmath0 relative to @xmath350 . much deeper is the question of what , exactly , is the large scale geometric behavior of these two spaces : the curve complex @xmath332 is a gromov hyperbolic metric space , and @xmath0 is weakly hyperbolic relative to stabilizers subgroups of representatives of orbits in @xmath23 . the boundary of the curve complex was identified in the following theorem of klarreich @xcite ; see also @xcite . two points in @xmath14 are _ topologically equivalent _ if they are represented by singular foliations which , after forgetting measure , are identical . [ theoremccboundary ] there is an @xmath0 equivariant quotient map from the subspace of @xmath14 consisting of classes of arational measured foliations to the gromov boundary of @xmath332 , such that the decomposition elements of this quotient map are the topological equivalence classes of arational measured foliations . using the fact that @xmath332 is convex cocompact , it makes sense to compare convex cocompactness of a subgroup of @xmath0 with respect to its actions on @xmath13 and on @xmath332 . this is accomplished in the following theorem proved independently , and with different techniques , by kent - leininger and hamenstadt : [ theoremcurvecomplex ] a finitely generated subgroup @xmath319 is convex cocompact if and only if its action on @xmath332 is convex cocompact . [ [ schottky - and - virtual - schottky - subgroups - of - mapping - class - groups . ] ] schottky and virtual schottky subgroups of mapping class groups . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a convex cocompact , free subgroup of @xmath0 is called a _ schottky subgroup_. the ubiquity of schottky subgroups was first demonstrated by farb and myself in @xcite , using a result of @xcite which is to be explained below . for any independent set of pseudo - anosov mapping classes @xmath355 , there exists an integer @xmath356 such that for all integers @xmath357 , the mapping classes @xmath358 freely generates a schottky subgroup of @xmath30 . kent and leininger @xcite and independently hamenstadt @xcite discovered a more direct and natural setting for this theorem , using a ping - pong argument in the curve complex . here is their proof . the action of @xmath359 on @xmath14 has source sink dynamics , with source and sink each being arational , and for @xmath360 the @xmath361 fixed points are pairwise topologically inequivalent , because they are pairwise distinct and each is uniquely ergodic . by theorem [ theoremccboundary ] , it follows that the action of each @xmath359 on @xmath362 has source sink dynamics . and so @xmath359 is a hyperbolic isometry on the gromov hyperbolic metric space @xmath332 , and that their fixed point sets in @xmath362 are pairwise disjoint . to conclude that @xmath358 freely generate a convex cocompact subgroup of @xmath0 , we apply the following folk theorem : if @xmath104 is a gromov hyperbolic geodesic metric space , and if @xmath363 are hyperbolic isometries of @xmath104 with pairwise disjoint fixed points in @xmath116 , then there exists an integer @xmath356 such that for all integers @xmath357 , the isometries @xmath358 freely generate a convex cocompact group of isometries . it seems hard to track down an early proof of hyperbolic ping - pong , although it was certainly known to the very earliest practitioners of gromov hyperbolicity . here is a sketch of a proof , pretty much the same as the proof given in @xcite . one may choose @xmath364 quasigeodesic lines @xmath365 in @xmath104 that are axes for @xmath363 , respectively ( if @xmath104 were proper then one could choose @xmath366 , @xmath367 , in other words , geodesics ) . fix a base point @xmath368 , and let @xmath369 be an upper bound for the distance from @xmath43 to @xmath365 . also , let @xmath370 be a lower bound for the length of a fundamental domain of each @xmath363 along its axis @xmath365 . fix @xmath371 and @xmath357 , let @xmath142 be the group generated by @xmath358 . we know by topological ping - pong that if @xmath371 is sufficiently large then @xmath142 is freely generated by @xmath358 . it suffices to show that for each word @xmath372 in the generators , a geodesic from @xmath43 to @xmath373 stays uniformly close to the orbit @xmath374 . we describe a path @xmath375 from @xmath43 to @xmath373 as follows . denote @xmath376 . starting from @xmath43 , jump a distance at most @xmath369 onto the axis @xmath377 , travel along @xmath378 for @xmath379 fundamental domains , thereby travelling a distance of at least @xmath380 , then jump a distance at most @xmath369 to the point @xmath381 . next , jump a distance at most @xmath369 onto the axis @xmath382 , travel along this axis for @xmath383 fundamental domains , thereby travelling a distance of at least @xmath380 , then jump a distance at most @xmath369 to the point @xmath384 . continuing in this manner we get a path @xmath375 from @xmath43 to @xmath373 , which stays in a @xmath385 neighborhood of @xmath374 . the path @xmath375 is a @xmath386 local @xmath387 quasigeodesic . if @xmath371 is sufficiently large , it follows that @xmath375 is a @xmath388 quasigeodesic for @xmath389 depending only on the previous constants and the hyperbolicity constant of @xmath104 @xcite . now we use the fact that @xmath375 has hausdorff distance at most @xmath390 from any geodesic @xmath200 with the same endpoints @xmath43 , @xmath373 , where @xmath390 depends only on @xmath134 , @xmath389 , and the hyperbolicity constant for @xmath104 . it follows that @xmath200 stays within distance @xmath391 of @xmath374 . [ [ virtual - schottky - subgroups . ] ] virtual schottky subgroups . + + + + + + + + + + + + + + + + + + + + + + + + + + + a _ virtual schottky subgroup _ of @xmath0 is a subgroup that contains a schottky subgroup with finite index . clearly convex cocompactness is preserved under passage to finite index subgroups , so every virtual schottky subgroup is convex cocompact . the following recent result , to appear in the dissertation of honglin min , is another example of a ping - pong argument in the mapping class group : if @xmath392 are finite subgroups of @xmath0 , if @xmath137 is pseudo - anosov , and if the virtual normalizer of @xmath127 has trivial intersection with @xmath235 and @xmath371 , then for all sufficiently large @xmath173 the subgroups @xmath235 and @xmath393 freely generate their free product in @xmath0 , and this subgroup is virtually schottky . [ [ hyperbolic - extensions - of - surface - groups ] ] hyperbolic extensions of surface groups + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + when the surface @xmath1 is closed and oriented , there is a canonical short exact sequence @xmath394 see @xcite . given a subgroup @xmath395 , let @xmath396 be its preimage in @xmath397 , and so we have an extension @xmath398 thurston s geometrization theorem ( see @xcite ) shows that if @xmath224 is an infinite cyclic , pseudo - anosov subgroup of @xmath30 then @xmath396 is a word hyperbolic group . the following theorem was the prequel to the theory of convex cocompact groups . it is proved by an application of the bestvina feighn combination theorem . if @xmath363 are independent pseudo - anosov elements of @xmath0 then sufficiently high powers of @xmath363 freely generate a subgroup @xmath224 such that @xmath396 is word hyperbolic . with this result as motivation , in @xcite farb and i set out to develop a geometric understanding of when @xmath396 is word hyperbolic , attempting to prove that for this to be true it was necessary and sufficient that @xmath224 be convex cocompact . we proved necessity , and when @xmath224 is free we proved sufficiency by using the bestvina feighn combination theorem @xcite . a completely general proof of sufficiency was recently given by hamenstadt @xcite , so that now one can state : [ theoremhypextension ] for any subgroup @xmath399 , the group @xmath396 is word hyperbolic if and only if @xmath224 is convex cocompact one very well studied class of subgroups of @xmath0 is stabilizers of teichmller discs . a _ teichmller disc _ in @xmath13 can be defined in one of the following equivalent ways . first , consider @xmath27 and @xmath92 . as @xmath400 varies over the unit circle in the complex plane , consider the family of quadratic differentials @xmath401 . this defines a family of rays based at @xmath29 parameterized by the circle , and the union of these rays is defined to be the _ teichmller disc with basepoint @xmath29 and direction @xmath33 _ , denoted @xmath402 . a somewhat more invariant way to say almost the same thing is that @xmath402 is the set of points in @xmath13 obtained by choosing an element of @xmath403 acting affinely on the euclidean plane , and applying that element to the regular canonical coordinates for @xmath33 to get a new conformal structure . second , a teichmller disc @xmath369 is a maximal subset which is the image of a holomorphic embedding of the poincar disc into @xmath13 . a theorem of royden @xcite in the compact case and gardiner @xcite in the noncompact case says that any two points @xmath404 lie on a unique such disc @xmath369 , and moreover that the teichmller distance between @xmath29 and @xmath53 equals their distance in the poincar metric on @xmath369 . the identification of @xmath369 with @xmath402 for some @xmath33 is obtained by choosing @xmath33 to be the initial quadratic differential from @xmath29 to @xmath53 . to each teichmller disc @xmath369 there is associated an embedded circle @xmath405 , defined as follows . choose @xmath27 and @xmath92 so that @xmath406 . as @xmath400 varies over the unique circle in the complex plane , the horizontal measured foliations of @xmath401 trace out the desired circle @xmath407 . alternatively , a point @xmath90 \in { \mathcal{pmf}}$ ] is in @xmath407 if and only if it is represented as a linear measured foliation in any regular canonical coordinate @xmath32 for @xmath33 ; in other words , there exists an @xmath408-linear differential form @xmath409 such that @xmath90 $ ] is represented in any @xmath32 by @xmath410 . this makes it clear that @xmath407 is well - defined independent of @xmath411 , because any other @xmath412 is obtained by a teichmller deformation with initial quadratic differential @xmath401 for some @xmath400 in the unit circle , resulting in a terminal quadratic differential @xmath52 on @xmath53 , and linearity with respect to @xmath33 and to @xmath52 are clearly equivalent . let @xmath413 be a teichmller disc , and let @xmath414 be its stabilizer . the group @xmath415 , or its image in @xmath403 when regarding @xmath369 as the upper half plane , is sometimes called a _ veech group _ , and the quotient orbifold @xmath416 , a proper holomorphic curve in the moduli space of @xmath1 , is sometimes called a _ veech surface_. these terms are sometimes restricted to the case where @xmath415 is a lattice , meaning that @xmath416 has finite hyperbolic area . regarding @xmath369 as the hyperbolic plane , the elliptic loxodromic trichotomy for @xmath417 becomes not just an analogy but a strict correspondence with the finite order pseudo - anosov dichotomy for elements of @xmath415 ; for details see see for example @xcite . to be precise , fix @xmath418 . then @xmath127 is loxodromic on @xmath369 if and only if it is pseudo - anosov in @xmath0 , @xmath127 is parabolic on @xmath369 if and only if it is reducible in @xmath0 , and @xmath127 is elliptic on @xmath369 if and only if it is finite order in @xmath0 . the case where @xmath127 is parabolic can be analyzed more closely . in this case , letting @xmath406 , the action of @xmath127 is a shear transformation on any regular canonical coordinate @xmath32 , parallel to the leaves of some constant slope measured foliation @xmath63 called the _ shear foliation _ for @xmath127 ; this foliation corresponds to the point of @xmath407 fixed by @xmath127 . each leaf of @xmath63 is compact , and so @xmath63 is obtained by enlarging some weighted family of simple closed curves @xmath419 . in particular , the essential curve family @xmath420 is a canonical reducing system for @xmath127 , and the first return of @xmath127 to each complementary component of @xmath421 is of finite order , so @xmath127 is the composition of a finite order mapping class that is reduced by @xmath420 and a product of dehn twists along the components of @xmath419 . using this description , we can get a fairly precise understanding of the case when @xmath415 is a reducible subgroup of @xmath0 . assuming this is so , let @xmath418 be of infinite order , with shear foliation @xmath63 . for any other infinite order @xmath422 , the shear foliation of @xmath246 must have the same slope as @xmath63 and so represent the same point in @xmath407 . otherwise , if the shear foliation @xmath97 of @xmath246 is distinct from @xmath63 in @xmath407 , then high powers of @xmath127 and @xmath246 would generate a loxodromic element of @xmath415 , by the proof of the tits alternative for @xmath20 ( this is not hyperbolic ping - pong , because @xmath127 and @xmath246 do not have source sink dynamics , but a variation of hyperbolic pong - pong works ) . thus , every infinite order element of @xmath415 stabilizes the same point in @xmath407 , in fact every such element acts as a shear transformation on @xmath33 with respect to the same linear measured foliation @xmath423 . but this implies that @xmath415 is virtually cyclic , generated by a multiple dehn twist about components of the curves @xmath419 whose enlargement is @xmath423 . in this case we can identify the mccarthy papadopoulos limit set of @xmath415 : it is the simplex in @xmath14 spanned by the projective classes of @xmath419 . henceforth we shall only be interested in the case when @xmath415 is infinite and irreducible . in this case , since @xmath407 is invariant under @xmath415 by construction , and since @xmath407 is obviously closed in @xmath14 , by applying theorem [ theoremmcpapairred ] we obtain : [ theoremveechlimits ] if @xmath415 is infinite and irreducible then the mccarthy papadopoulos limit set of @xmath415 is contained in @xmath407 . thurston @xcite gave the first examples of teichmller discs @xmath369 such that @xmath415 is a lattice , meaning that it acts on @xmath369 with cofinite area . start with the linear action of @xmath150 on @xmath424 , then choose a finite subset @xmath425 invariant under this action , and let @xmath426 be a branched cover of @xmath427 branched over @xmath235 . lifting the conformal structure from @xmath427 to @xmath1 defines @xmath428 , and lifting the quadratic differential @xmath429 from @xmath427 to @xmath1 defines @xmath92 . there exists a finite index subgroup @xmath430 that lifts to @xmath1 , and @xmath142 stabilizes @xmath402 and acts with cofinite area . combining with theorem [ theoremveechlimits ] it follows that the mccarthy papadopoulos limit set of @xmath142 is the circle @xmath407 . veech @xcite gave the first examples where @xmath415 is a lattice that does not arise from thurston s construction of branching over the torus . veech also developed some of the general theory of @xmath415 , for instance establishing that @xmath415 never acts cocompactly on @xmath369 . when @xmath415 is a lattice , it follows that there are finitely many maximal parabolic subgroups of @xmath415 up to conjugacy , and the veech surface @xmath416 is a finite area holomorphic curve in the moduli space of @xmath1 . mcmullen @xcite proved that if @xmath1 is the closed surface of genus 2 , if @xmath431 is a teichmller disc , and if @xmath415 contains a hyperbolic element , then the limit set of @xmath415 is all of @xmath407 . in the same paper mcmullen also gave examples where @xmath415 is infinitely generated , a phenomenon independently discovered by hubert and schmidt @xcite . for a further survey of stabilizers of teichmller discs , and references to the literature , see the section `` veech surfaces '' in the problem list @xcite , and the chapters by harvey @xcite and by herrlich and schmithsen @xcite in this handbook . the only known examples of convex cocompact subgroups are free groups and virtually free groups . even in @xmath19 , constructing nonfree discrete subgroups can be subtle . one of the pioneering tools in this regard is the maskit combination theorem @xcite . one of the most interesting questions about subgroups of mapping class groups is whether there exists closed surfaces @xmath432 of genus @xmath18 and a convex cocompact subgroup of @xmath30 isomorphic to @xmath433 . if such a subgroup existed then , by hamenstadt s theorem , one would obtain the first example of a compact 4-manifold @xmath434 with word hyperbolic fundamental group that fibers over a surface . the preprint @xcite of m. kapovich contains a theorem that says if such an @xmath434 existed then @xmath434 could not have a @xmath435 structure . it is even hard to construct any surface subgroup @xmath433 of @xmath30 with the genus of @xmath436 at least two . the first examples were given by harvey and gonzlez - dez @xcite , but they have many non - pseudo - anosov elements and so can not be convex cocompact . the starting point of the theorem is the construction of a teichmller disc @xmath369 and finite index subgroup @xmath437 such that @xmath415 ( and hence also @xmath142 ) is a lattice , @xmath142 is torsion free , and @xmath142 has exactly one parabolic subgroup , generated by a multitwist @xmath438 about a multicurve @xmath21 such that no component of @xmath71 is a three holed sphere . it follows that there exists a homeomorphism @xmath198 that preserves @xmath21 and is pseudo - anosov on each component of @xmath71 . with the objects as denoted above , for each sufficiently large @xmath41 the group generated by @xmath142 and @xmath439 is a free product with amalgamation along @xmath440 , and is therefore isomorphic to @xmath433 for an oriented surface @xmath436 of genus @xmath18 . moreover , every element of @xmath441 is pseudo - anosov _ except _ for elements conjugate to powers of @xmath438 . a leininger reid subgroup can not be convex cocompact , because it contains a multitwist . in @xcite , a notion of geometric finiteness for subgroups of @xmath0 is proposed , and the question is posed whether leininger reid subgroups are geometrically finite . f. herrlich and g. schmithsen , _ on the boundary of teichmller disks in teichmller space and in schottky space _ , handbook of teichmller theory , volume 1 ( a. papadopolous , ed . ) , european mathematical society , 2007 , to appear . serre , _ rigidit du foncteur de jacobi dchelon @xmath442 _ , appendix to a. grothendieck , _ techniques de construction en gomtrie analytique , x. construction de lespace de teichmller _ , sminaire henri cartan , no . 17 , secrtariat mathmatique de lens , paris , 1960/61 . w. a. veech , _ teichmller curves in moduli space , eisenstein series and an application to triangular billiards _ , invent . math . * 97 * ( 1989 ) , no . 3 , 553583 , erratum : volume 103 number 2 , 1991 , page 447 .
the theme of this survey is that subgroups of the mapping class group @xmath0 of a finite type surface @xmath1 can be studied via the geometric / dynamical properties of their action on the thurston compactification of teichmller space @xmath2 , just as discrete subgroups of the isometries of hyperbolic space @xmath3 can be studied via their action on compactified hyperbolic space @xmath4 . 30f60 , 57m50 mapping class group
one of the most important findings of nasa s mission is the prevalence of planets on relatively short period orbits . m dwarfs host , on average , 2.5 planets per star with radii between 1 and 4@xmath1 and orbital periods shorter than 200 days @xcite and nearly half of all g and k stars host at least 1 planet within the same range @xcite . the ubiquity of these planetary systems has sparked a number of investigations of their underlying architecture . the most straight - forward observable to compare with predictions is the number of transiting planets ( tranets " ) per system . constructing a set of synthetic planetary systems and then observing " them from random angles provides a means of comparison to the tranet yield . these studies constrain the underlying multiplicity of the systems and distribution of planet inclinations . the general consensus from these simulations is that planetary systems tend to be very flat ( mutual inclinations less than a few degrees ) @xcite . however , there exists a strong degeneracy between the underlying multiplicity and the mutual inclinations of the planets . that is , systems with more planets with higher mutual inclinations will present the same number of transiting planets as systems with fewer planets with lower mutual inclinations . transit duration ratios of adjacent planets , which are sensitive to the inclination distribution of planets , provide a means to break this degeneracy . both @xcite and @xcite folded transit duration ratios into their analyses and concluded that most systems have 1 - 2 planets with mutual inclinations between planets less than 3@xmath2 . other studies have found that the observed population of planets is not well described by a single - component model of system architecture . @xcite and @xcite both find that their best fitting single component models significantly under - predict the number of singly transiting systems . @xcite compare planetary systems created in late stage planet formation simulations to observed systems and similarly find that singly transiting systems are 50% more common in the observed population than their modeled population . looking beyond multiplicity , @xcite compared the fraction of transiting planets showing transit timing variations ( ttvs ) for single vs multi - planet systems and found that systems with four or more transiting planets show ttvs at a rate @xmath35 times higher than those of singly transiting systems . these findings all suggest that the underlying architecture of planetary systems is better described by either a mixture model or a continuum of architectures . qualitatively , a mixture model must consist of one component with many planets on low mutual inclinations that accounts for the majority of systems with two or more transiting planets . the other component must account for the majority of systems with only one transiting planet . these latter systems either host fewer planets , or their planets possess high mutual inclinations so that only one planet will transit from any given viewing angle . the utility of transit duration ratios as a tool for probing mutual inclination is lost for systems with only one transit . @xcite find that the stellar obliquity of systems with only one transiting planet are systematically higher than for systems with multiple transiting planets . this may suggest that the second population is made up of planets with high mutual inclinations . however , it is possible that whichever potential mechanism reduces the underlying multiplicity of systems also increases obliquity . a number of authors have explored the possibility that a dichotomy in the structure of planetary systems can arise from the long term evolution of dynamically full systems . @xcite suggest that systems of many tightly packed planets are the normal outcome of planet formation and that some fraction of these systems eventually undergo a dynamical instability which leads to the consolidation of these systems into fewer planets . similarly , @xcite find that many of the multi - planet systems are metastable . they propose that the lower multiplicity systems are resultant from more densely packed systems that have undergone a period of dynamical instability . on the other hand , @xcite find that dynamical instability is probably not the cause of the dichotomy . they followed the evolution of triple systems with the goal of producing doubles with high mutual inclinations . however , they found that the timescales needed for instability to occur are very long unless the planets are very high mass . planets of 0.1 to 10 jupiter masses are needed for instability to occur within 10 billion years . similarly @xcite found that self - excitation is rarely sufficient to remove planets from the transiting geometry . we explore this hypothesis through n - body simulations of late stage planet formation and a comparison of these simulation results to the observed population of planets . we investigate whether two properties : ( 1 ) the total mass of planetesimals within the protoplanetary disk and ( 2 ) the radial distribution of this mass determine the final multiplicity of planetary systems and the mutual orbital inclinations of the planets . in section 2 , we describe the details of our simulations , and in section 3 describe their results . in section 4 , we compare observable quantities from our simulations : number of transiting planets , transit duration ratio , and period ratio , to observations . in section 5 , we describe the application of our simulations specifically to the sample of planets orbiting m dwarfs ( @xmath45.1 ) . we go on to comment upon the possible physical origin of the variation in the properties of planetesimal disks ( @xmath45.2 ) and the long - term evolution of planetary systems ( @xmath45.3 ) . in @xmath45.4 , we summarize the predictions from our simulations and implications for future study , and conclude in section 6 . the formation pathway of short period planets is a matter of debate . models of their formation fall into two general categories : those where planets form in situ from a massive planetesimal disk ( e.g. * ? ? ? * ; * ? ? ? * ) and those where planets form at larger distances from the star and later migrate to their currently observed locations ( e.g. * ? ? ? * ; * ? ? ? currently , comparison with known planetary systems does not definitively favor one model over the other . although both models are able to form systems that generally match the architectures of observed planetary systems , neither of them have yet been able to duplicate all of their characteristics . it is expected that bodies larger than about the mass of mars should feel significant torques from the disk @xcite ; enough for substantial migration over the lifetime of a protoplanetary disk . on the other hand , a strong dependence on detailed physical structure of the disk leads to uncertainties in the magnitude and even direction of migration @xcite indicating we may not yet know enough about planet migration to definitively conclude that it is important for the formation of close - in planets . in this work , we assume that planets form in situ from a previously assembled planetesimal disk . in order to evaluate this hypothesis , we ran a large set of n - body simulations of late - stage planet formation . we varied both the total mass of the planetesimal disk and the radial distribution of this mass . the initial surface density of the planetesimal disk is assumed to follow a power law : @xmath5 we choose @xmath6 to be -2.5 , -1.5 , -0.5 and set @xmath7 so that the total mass between 0.05 and 1 au is 7 , 10 , 17 or 35 @xmath8 . for each combination of @xmath6 and disk mass we ran 32 simulations with different realizations of the initial planetesimal disk . motivated by the results of @xcite , we break the initial planetesimal disk into two populations : larger bodies , which we refer to as planetary embryos , and smaller bodies , which we refer to as planetesimals . each type of body cumulatively accounts for 50% of the total planetesimal disk mass . the initial masses and semi - major axes of the embryos are determined by enforcing two requirements : 1 . their radial mass distribution follows the chosen surface density power law profile ( divided by two since they only account for half the mass ) and 2 . they are separated by about 15 mutual hill radii . depending on the surface density profile , this results in anywhere from 15 to 65 embryos with a range of masses ( which scale with semi - major axis as described by equation [ expectedmass ] ) . for each embryo , we place ten planetesimals with a tenth of the embryo s mass randomly positioned half way between the orbits of the embryo s inner and outer - most nearest neighbors . the initial orbits of all bodies are close to circular ( @xmath9 ) and have low inclinations ( @xmath10 ) . the remaining orbital elements are assigned random values between 0 and 360 degrees . the evolution of each planetesimal disk around a solar mass star was integrated using the mercury n - body integrator @xcite . collisions are assumed to result in perfect mergers of the bodies . the systems were integrated for 10 million years with a timestep of 0.5 days . we begin by constructing a simple model to predict the masses of planets as a function of semi - major axis given a set of initial disk conditions . in this model , the mass of a planet is determined by the size of its feeding zone and the amount of mass in that zone . we assume that the size of the feeding zone is similar to the distance between the planet and its neighbor and that this separation is some number , @xmath11 , times their mutual hill radius , @xmath12 given an initial power law surface density distribution and a separation between neighboring planets , @xmath11 , the expected mass of a planet forming at a given semi - major axis is : @xmath13 figure [ hillseparations ] shows the expected range of planetary masses ( shaded in grey ) assuming @xmath14 ( values typical for planetary systems formed in n - body simulations ) for each different initial planetesimal surface density distribution compared to the actual mass of planets that formed in the simulations . for @xmath15 , this simplistic model predicts , reasonably well , the masses of planets . for larger @xmath6 , the model is a poor predictor of planet mass . the failure of this model is not surprising considering that it ignores the complex dynamical interactions present in these systems . here , we highlight two factors that influence the mass and spacing of planets : the radial redistribution of mass and the orbital eccentricity . equation [ expectedmass ] is only valid if the final mass surface density distribution is the same as the initial one used to derive the equation . however , the numerous gravitational scatterings and resonant interactions that occur in the n - body simulations can significantly redistribute mass in the system . figure [ redistribution ] shows the averaged surface density distributions of the final planetary systems relative to the initial planetesimal surface density distributions . it is evident that there is some amount of mass redistribution in all systems . in particular , the outer disk , beyond @xmath30.6 au loses mass during the simulation . some of this mass is scattered outward and to conserve energy some must move inward . the inward movement of material is more apparent in the shallower surface density profile disks because there is less mass in the inner disk to begin with and any addition to this is more noticeable . this decrease of mass in the outer disk is reflected in figure [ massvssma ] in which the simulated planets are undermassive compared to what is expected from equation [ expectedmass ] . the eccentricity affects the final mass of a planet because it changes the width of that planet s feeding zone . a planet on an eccentric orbit will occupy a larger radial range of the disk than one on a circular orbit . this effectively increases the value of @xmath11 in equation [ expectedmass ] leading to a larger expected mass . this affect can be seen in figures [ hillseparations ] and [ eccentricityvssma ] . for @xmath6 = -1.5 and -0.5 , in the inner disk there is a rise in eccentricity and a corresponding rise in the orbital separations between neighboring planets . thus we would expect for this region that the masses of planets that form in the simulations are larger than our simple prediction from equation [ expectedmass ] . indeed , this is what we see in figure [ massvssma ] . we can start to understand the trends in eccentricity and planet - planet separations with semi - major axis by considering the angular momentum of these systems , or alternatively , the angular momentum deficit ( amd ) . the amd is the difference in angular momentum that a planetary system has compared to what it would have if the orbits were circular and coplanar : @xmath16 where @xmath17 is the circular angular momentum of the @xmath18th planet . the amd of a system provides some measure of its stability . a system with zero amd ( composed of coplanar and circular orbits ) will remain stable forever , whereas a system with sufficiently high amd will exhibit secular chaos @xcite . secular chaos enables the transfer of amd between planets leading to potentially large values of their eccentricity and inclination . because amd is conserved in secularly interacting systems , the maximum eccentricity a planet can attain can be calculated from equation [ amd ] by assuming the planet accounts for the entirety of the amd for the system ( i.e. the other planets are on circular and coplanar orbits ) . this is not the case for systems with resonant interactions , as in many of our simulations . however the amd varies by no more than a factor of a few after the first million years for most of our simulations and so the maximum eccentricity calculation is still approximately true . figure [ maxeccentricity ] shows the maximum eccentricity calculated for each of our simulated planets given the amd , masses and semi - major axes of the planets at the end of the simulation . the same general trends of eccentricity with semi - major axis and initial planetesimal disk structure are apparent in the calculated maximum eccentricities and the actual eccentricities albeit with an offset in eccentricity , indicating that secular chaos is a likely cause for the trends in eccentricity . for the simulations with @xmath19 planet mass increases with semi - major axis . accordingly , the angular momentum of the inner planets is much smaller than that of the outer planets . in this case , a moderate fraction of the total amd of the system can be large compared to the angular momentum of one of the inner planets . thus if this amd is transferred to that planet , its eccentricity can become quite large . on the other hand , in the simulations with @xmath15 , masses decrease with semi - major axis . the angular momentum of individual planets in these systems are more similar and accordingly their maximum eccentricities do not vary significantly with semi - major axis . the distribution of masses and orbital separations of planets ultimately determine the final multiplicity of the planetary systems . given some mass budget , an increase in orbital separation and the corresponding increase in planet mass will result in fewer planets in the system . because the systems formed in simulations with larger values of @xmath6 form , on average , more massive and more widely separated planets , we expect their multiplicity to be lower . this is demonstrated in figure [ intrinsicmultiplicity ] . changing the initial distribution of mass in the planetesimal disk leads to a significant change in the multiplicity of systems that form . larger @xmath6 and higher disk mass correspond to the formation of significantly fewer planets . as we will show in section [ comparison ] , this has a strong impact on the observed multiplicity of planetary systems . the other determining factor in the observed multiplicity of systems is the spread in mutual inclinations between planets . the effects of the initial planetesimal disk conditions on final planet inclination are significant . more massive disks tend to produce planets with moderately higher mutual inclinations and disks with larger @xmath6 produce planets with significantly larger mutual inclinations ( see figure [ inclinations ] ) . in order to compare our simulated planetary systems to the catalog of known exoplanets , we observe " the final state of each dynamical simulation 10,000 times . we simulate the viewing geometry isotropically . our final synthetic yield of transiting planets are those ( 1 ) with impact parameter @xmath20 ( from the given viewing angle ) and ( 2 ) whose probability of detection by exceeds a randomly drawn uniform number between 0 and 1 . we first compare our simulations to observations of _ kepler _ transiting planets orbiting gk stars . to create a realistic synthetic sample to compare to the observed sample of transiting planets , we employ the pipeline completeness calculation . @xcite contains a likelihood of detection as a function of orbital period and planetary radius . this model uses the completeness parameterization presented in @xcite , itself generated from signal injection and recovery in data . because our simulations calculate planet mass , we translate the planetary masses to radii to determine a detection probability for each planet . we use the probabilistic relationship presented in @xcite for @xmath21 ( corresponding to @xmath22 , with the restriction that the radius must be greater than 1.2 @xmath23 . for planets with masses less than @xmath24 , we use the best fit deterministic model from @xcite because the data does not rule out a deterministic mass / radius relation for small planets and the scatter in the probabilistic model is unphysical for small planets . our comparison sample consists of all known planets orbiting the stars used to calculate the detection probability grid . this sample is made up of 2,653 planets in 2042 systems . figure [ observedmultiplicity ] shows the distribution of apparent multiplicities from our set of observed " systems for each initial disk structure . there is a clear dependence of the observed multiplicity on the initial planetesimal disk distribution , showing variations with both the surface density power law index , @xmath6 , and the total initial mass in the disk , @xmath25 . our observed data set @xmath26 is a vector with six entries , corresponding to the number of stars hosting @xmath27 tranets , where 1@xmath286 . in our final sample , there are 1635 g and k dwarfs hosting 1 detected transiting planet , 269 hosting 2 , 88 hosting 3 , 36 hosting 4 , 12 hosting 5 , and 2 hosting 6 . we compare the model predicted population , @xmath29 ( with results shown in figure [ observedmultiplicity ] ) to the observed yield @xmath30 of tranets . we estimate the likelihood of the model we have generated from our simulations , @xmath29 , given @xmath30 assuming the number of systems with @xmath27 transiting planets is described by poisson counting statistics . therefore , we can describe the likelihood of the model as : @xmath31 where @xmath32 from our dynamical simulations , and @xmath27 runs from 1 to 6 . we first evaluate @xmath33 using only one mode of planet formation : that is , assuming one value for @xmath6 and @xmath25 . our grid of @xmath6 and @xmath25 consists of 12 points , with 3 possible values of @xmath6 ( -0.5 , -1.5 , and -2.5 ) and 4 possible values of @xmath25 ( 7 , 10 , 17 , 35 @xmath8 ) . the model distribution , @xmath29 , is created by re - normalizing the simulated multiplicity distribution ( shown in [ observedmultiplicity ] ) to the number of systems in our observed sample . we also employ a mixture model , to test whether the data support a model with a range of initial conditions for planet formation . we assume the simplest case of two modes , each with independent @xmath6 and @xmath25 , occurring with respective frequency @xmath34 and @xmath35 . the model distribution , @xmath29 , is the weighted sum of the multiplicity distributions of each mode . we assign a flat prior on both @xmath6 and @xmath25 , and employ the bayesian sampler multinest @xcite to evaluate the posterior distributions . we compare the output number of tranets from our simulations to the observed tranets of the gk host stars . in agreement with previous studies , we find that the data favor a two - component model over a single component model . taking the ratio of the bayesian evidences returned by multinest , we find a log evidence for a mixture model of -8.8 , and log evidence of -26.7 for a single mode model . we conclude that the data are better supported by a mixture model by a factor of 18:1 . in figure [ fig : tranets_fit ] , we see that the high multiplicity tail of the multiplicity distribution requires one of the two modes to correspond to systems formed in disks with the steepest surface density profile ( @xmath6=-2.5 ) and a disk mass of 17 earth - masses or greater . the mode that accounts for the high number of singly transiting systems is less well constrained because many of the disk configurations lead to systems that typically show only one transiting planet . however , models with shallower surface density profiles are favored . we find that between 15 and 35% of systems must form in the first mode while the rest form in the second . this finding resembles that of @xcite who find that 75 - 80% of systems account for the majority of low observed multiplicity systems . however , our conclusions regarding the underlying architecture of these systems differ . they find that most systems are made up of 1 - 2 planets with low mutual inclinations , whereas we find the underlying multiplicity and mutual inclinations to be substantially larger . studies such as @xcite , @xcite , and @xcite employ ensemble normalized transit durations in multi - tranet systems to infer average orbital mutual inclination . transit duration depends upon orbital period as well as impact parameter . the ratio of transit durations , @xmath36 ( normalized by orbital velocity , per equation [ eq : duration ] ) , of adjacent planets therefore encodes information about their mutual inclination . @xmath37 in figure [ fig : perratio_fit ] , we show the observed distribution of orbit - normalized transit durations and the observed distribution of period ratios for transiting planets ( both in blue ) . we overplot our model _ predictions _ for these distributions that we infer from our multiplicity fit . we see that the two - component model self - consistently recovers not only the tranet distribution , but also the duration and period ratio distributions ( neither of which are well - approximated by a single model ) . we also examine the effect of stellar mass upon our simulations . there exists suggestive evidence that planet formation proceeds differently around m dwarfs than around gk dwarfs . the relationship between stellar metallicity and planet occurrence may vary as a function of stellar mass @xcite , as well as the mass distribution of the planets themselves @xcite . m dwarfs host , on average , 2.5 planets per star , many more than sun - like stars @xcite . intriguingly , both gk dwarfs @xcite and m dwarfs @xcite also exhibit a dichotomy in their underlying architecture . for m dwarfs , a mixture model best recovers the observed multiplicity distribution . half of these systems contain at least 5 planets with low mutual inclinations , and the other half containing either one planet or multiple planets with high mutual inclinations . the difference in multiplicity distribution between gk dwarfs and m dwarfs is also suggestive : 1/3rd of m dwarfs planet hosts uncovered by host 2 or more transiting planets , while this figure is 1 in 5 for gk dwarfs . however , s sensitivity to transits varies with stellar mass . a meaningful comparison requires folding s completeness to planets into our multiplicity study . to address this , we performed a separate set of simulations with a stellar mass of 0.5 @xmath38 . this value corresponds to the mass of the typical m dwarf observed by . for this set of simulations , we began with planetesimal disks extending from 0.05 - 0.5 au . rather than varying total mass in the disk as with our gk dwarf simulations , we make the simplifying assumption that the disk contains 10 @xmath8 of material . we varied the surface density profile by changing the power law index to be -2.5 , -1.5 , -0.5 , and 0.0 . for each value of the power law index we ran eight simulations . for comparison with the sample , we generated synthetic observations " in the same way we did for the solar - mass sample . we employed the detection efficiency grid for planets around m dwarfs presented in @xcite . we used the sample of planets detected in that work to ensure the detection efficiency was the same for the observed sample and our simulated sample . the results of the tranet fit are shown in figure [ fig : tranets_mstars_fit ] . we find that the data favor a mixture model by a more modest value for m dwarfs than for the gk dwarfs . a mixture model in this case returns a log evidence of -8.4 , while a single mode model returns -10.5 . the 2:1 evidence ratio we find in this work is less than 8:1 stated by @xcite . we attribute this difference to the relative poorness - of - fit of even the mixture model in this case . because we limited our suite of m dwarf simulations to those with a total mass of 10@xmath8 , we can not well reproduce the high tranet multiplicity that we observe . within our gk simulations , we found that we required total masses of 17 @xmath8 to recover the shape of the multiplicity distribution . we infer that increasing the available mass in our set of m dwarf simulations would similarly result in a better fit to observations . the fractions @xmath34 of m dwarf planetary systems that reside in the flatter geometry is 63@xmath016% , versus the 24@xmath07% of gk planetary systems . this value is consistent with the value of @xmath34 from @xcite for m dwarf planetary systems of 47@xmath010% . the hypothesis of two distinct values for @xmath34 is favored by 611:1 , as compared to the hypothesis that these posteriors are drawn from the same underlying distribution . we conclude that there is strong evidence for a different mixture of primordial disks around gk stars than around m dwarfs . direct observational constraints on the distribution of mass in planetesimal disks are lacking . estimates of planetesimal disk structure can be made by comparison to the mass profile of dust and gas disks ( e.g. * ? ? ? * ) or planetary systems ( e.g. * ? ? ? * ; * ? ? ? however , neither the dust and gas disks nor the final systems that form from them necessarily have the same surface density profile as their planetesimal disk . planetesimal disks form within the gas / dust disk , but they are not coupled to its viscous evolution . consequently , the evolution of their structure is distinct . once the planetesimal disk forms , the final planetary system is built through the scattering and merger of planetesimals . these scattering events can lead to a significant redistribution of mass in the system thus erasing any direct connection between the surface density profiles of the initial and final states . the structure of planetesimal disks can also be approached from a modeling standpoint . @xcite simulated the growth of planetesimal disks from the accretion of inward drifting pebbles . they found that variations in the mass of the protoplanetary disk , disk metallicity or the duration of accretion lead to planetesimal disks with different masses and radial profiles . thus the range of initial planetesimal disk structures required for this work may be a natural consequence of the expected variation among protoplanetary disks . the growth of planetesimal disks into planetary systems is a relatively fast process in the inner disk . most of the evolution of these systems occurs within the first million years with collision rates dropping off significantly after that . that is not to say that longer - term evolution of these systems is not important in determining the final state of planetary systems . other works ( e.g. * ? ? ? * ; * ? ? ? * ) have studied how late - term dynamical instabilities can potentially reduce the number of planets in a system . the applicability of these results to the systems formed in our simulations is unclear because the initial structure of our planetary systems ( e.g. planet masses and spacings ) differs from theirs . we extended a subset of 32 simulations out to about 5 billion years . of the simulations that we extended , the observed planet multiplicity distribution did not change significantly between 10 million and 5 billion years . it is possible that a larger sample would result in more dynamical instabilities but unlikely at a high enough rate to account for the number of singly - transiting systems . tides are generally accepted to be an important factor in the orbital evolution of planets , particularly at small semi - major axes . observations of exoplanets show an overall decrease in orbital eccentricity the closer the planet is to the star , which is attributed to tidal circularization . tides would not have a large effect on our simulations because the timescale for tidal evolution is generally long compared to the duration of our simulations @xcite . however , they would have a significant impact on the subsequent evolution of the orbits of the innermost planets in our simulations . in particular , the omission of tidal physics likely explains the increase in eccentricity with decreasing semi - major axis seen in some of our simulations , which is not observed . the tidal circularization of orbits is also accompanied by an inward migration of order 0.01 - 0.05 au @xcite . both of these factors affect transit probability . inward migration will increase the transit probability . however , orbital circularization will counteract this increase to some extent because eccentric orbits have a higher transit probability @xcite . we note that the statistical comparisons made are not intended to ascertain the exact combination of initial conditions that must exist in planetesimal disks in order to produce the observed planetary systems . rather , they are meant to emphasize the inability of models that assume all planetary systems have the same underlying architecture to reproduce the observations and , furthermore , to show that the range of system architectures that result from differing initial conditions in the planetesimal disk can reproduce important aspects of observations . nevertheless , we can draw some important conclusions from our analysis . to begin with , we , like other works ( e.g. * ? ? ? * ) , find that a model for the final assembly of close - in planets that relies solely on the gravitational interactions of the growing bodies without the need for additional physics ( e.g. planet disk interactions ) can reproduce many of the observed characteristics of exoplanetary systems . assuming that close - in planetary systems are predominantly gravitationally assembled in this way , we can say that some variation in the initial planetesimal disk structure is needed to form the variety of observed architectures ( assuming there is no further relevant evolution of the systems after formation ) . in particular , some fraction of planetesimal disks must have initially steep radial surface density profiles while the rest have shallower surface density profiles . the planetary systems that form from this range of initial conditions have significantly different final architectures . because the systems that form from shallower surface density profiles typically only show at most one transiting planet and because these systems account for the majority of single tranet systems , we can make some general testable predictions about single tranet systems vs. multiple tranet systems : * single - tranet systems contain fewer , more widely separated ( in terms of mutual hill radii ) planets than do multiple - tranet systems . although not directly observable , this may be inferred by the frequency and strength of transit timing variations . * single - tranet systems should show larger stellar obliquity . this phenomenon is observed by both @xcite and @xcite . * planet mass should increase with semi - major axis in single - tranet systems whereas in multiple tranet systems it should decrease with semi - major axis . * single - tranet systems should have higher average eccentricity . this prediction is borne out in part by @xcite . they reported that average orbital eccentricity is a strong function of planetary multiplicity , with higher average eccentricity corresponding to fewer planets . additionally , because the diversity in architectures is determined at formation and is not the consequence of long - term evolution ( e.g. dynamical instability ) , we predict that there should be no correlation in a system s multiplicity and its age . in figure [ fig : hill_separation ] , we show our predicted distribution in the hill separation of _ detected _ planets that we infer from the tranet comparison . the bulk of planet pairs have separations between @xmath310 and 40 mutual hill radii . @xcite find that the distribution of separations for high multiplicity systems ( 4 + ) tends to be condensed into the 10 - 25 range , slightly smaller values than our overall distribution . however , the component of our distribution corresponding to the majority of multi - tranet systems and all high multiplicity systems ( @xmath6=-2.5 ) tends to produce planet pairs with smaller separations more consistent with the @xcite distribution ( see figure [ fig : hill_separation]b ) . [ fig : hill_separation ] recent analyses of exoplanet statistics have revealed that the configurations of planetary systems are not well described by a single - component model . models that account for the frequency of multi - tranet systems significantly under - predict the number of single - tranet systems and models that can account for the number of single - tranet systems under - predict the number of multi - tranet systems . we have investigated the hypothesis that these differences in system architecture arise from variation in the primordial distribution of mass in planetesimal disks . we performed a set of n - body simulations of late - stage planet formation , within which we varied initial planetesimal disk profiles . the resulting planetary systems are highly dependent on the initial conditions . multiplicity , mutual inclinations , orbital separations , and eccentricities of planets vary significantly as a function of both total planetesimal disk mass and the radial distribution of this mass . we compared our simulated systems to the yield of transiting planets . we apply the empirically - determined completeness to our synthetic populations of planets . in this way , we compare the outcome of our simulations to the observed multiplicities from . we conclude , similarly to previous works , that we can not replicate the yield of exoplanets with a single model for planet occurrence . rather , the data favor a mixture model , with at least two sets of distinct initial conditions required to replicate the observations from . we find that the same mixture model that correctly predicts the tranet distribution from also self - consistently reproduces the period ratio and duration ratio distributions . we find that the best fit underlying population to our gk dwarf sample consists of one component accounting for approximately 20% of systems that forms from a very steep initial planetesimal disk . this component accounts for most of the multi - tranet systems . the other component ( accounting for approximately 80% of systems ) forms from planetesimal disks with shallower surface density profiles . applying this analysis to m dwarfs , we see that the fraction of systems forming from steep planetesimal disk profiles increases to approximately 70% . we conclude that the differing nature of the dichotomy between gk and m dwarfs indicates that the process of planet formation varies with stellar mass . this material is , in part , based upon work supported by nasa under award no . nnx12ac01 g and performed , in part , under contract with the california institute of technology ( caltech ) funded by nasa through the sagan fellowship program , and also by the mit torres fellowship for exoplanet research . it is based on observations made with , which was competitively selected as the tenth discovery mission . funding for this mission is provided by nasa s science mission directorate . the authors would like to thank the many people who generously gave so much their time to make this mission a success . this research used the nasa exoplanet archive , which is operated by the california institute of technology , under contract with the national aeronautics and space administration under the exoplanet exploration program . the simulations for this work were supported by the facilities and staff of the yale university faculty of arts and sciences high performance computing center , and by the national science foundation under grant # cns 08 - 21132 that partially funded acquisition of the facilities . lastly , we thank eric agol , rory barnes , tom quinn and many others for their valuable input and advice .
nasa s mission uncovered a wealth of planetary systems , many with planets on short - period orbits . these short - period systems reside around 50% of sun - like stars and are similarly prevalent around m dwarfs . their formation and subsequent evolution is the subject of active debate . in this paper , we simulate late - stage , in - situ planet formation across a grid of planetesimal disks with varying surface density profiles and total mass . we compare simulation results with observable characteristics of the sample . we identify mixture models with different primordial planetesimal disk properties that self - consistently recover the multiplicity , period ratio and duration ratio distributions of the planets . we draw three main conclusions : ( 1 ) we favor a frozen - in " narrative for systems of short period planets , in which they are stable over long timescales , as opposed to metastable . ( 2 ) the dichotomy " , an observed phenomenon of the sample wherein the architectures of planetary systems appear to either vary significantly or have multiple modes , can naturally be explained by formation within planetesimal disks with varying surface density profiles . finally , ( 3 ) we quantify the nature of the dichotomy " for both gk stars and m dwarfs , and find that it varies with stellar type . while the mode of planet formation that accounts for highly multiplistic systems occurs in 24@xmath07% of planetary systems orbiting gk stars , it occurs in 63@xmath016% of planetary systems orbiting m dwarfs .
black hole entropy is given by a mysterious formula called the bekenstein - hawking formula @xcite : @xmath0 where @xmath1 is area of the horizon . there have been many attempts to understand the statistical origin of the black - hole entropy . entanglement entropy @xcite is one of the strongest candidates of the origin of black hole entropy . it is originated from a direct - sum structure of a hilbert space of a quantum system : for an element @xmath2 of the hilbert space @xmath3 of the form @xmath4 the entanglement entropy @xmath5 is defined by @xmath6,\nonumber\\ \rho_i & = & { \bf tr}_{ii}|\psi\rangle\langle\psi|.\end{aligned}\ ] ] here @xmath7 denotes a tensor product followed by a suitable completion and @xmath8 denotes a partial trace over @xmath9 , respectively . on the other hand , there is another strong candidate for the origin of black hole entropy : the brick wall model introduced by thooft @xcite . in this model , thermal atmosphere in equilibrium with a black hole is considered . in this situation , we encounter with two kinds of divergences in physical quantities . the first is due to infinite volume of the system and the second is due to infinite blue shift near the horizon . we are not interested in the first since it represents contribution from matter in the far distance . hence we introduce an outer boundary in order to make our system finite . it is the second divergence that we would like to associate with black hole entropy . namely , it can be shown by introducing a planck scale cutoff that entropy of the thermal atmosphere near the horizon is proportional to the area of the horizon in planck units . in this paper , we show that the brick wall model seeks the maximal value of the entanglement entropy . for simplicity , we consider a minimally coupled , real scalar field described by the action @xmath10,\ ] ] in the spherically symmetric , static black - hole spacetime @xmath11 we denote the area radius of the horizon by @xmath12 and the surface gravity by @xmath13 ( @xmath14 ) : @xmath15 we quantize the system of the scalar field with respect to the killing time @xmath16 in a kruskal - like extension of the black hole spacetime . the corresponding ground state is called the boulware state and its energy density is known to diverge near the horizon . although we shall only consider states with bounded energy density , it is convenient to express these states as excited states above the boulware ground state for technical reasons . hence , we would like to introduce an ultraviolet cutoff @xmath17 with dimension of length to control the divergence . the cutoff parameter @xmath17 is implemented so that we only consider two regions satisfying @xmath18 ( shaded regions @xmath19 and @xmath20 in _ figure _ [ fig : kruskal ] ) , where @xmath21 ( @xmath22 ) is determined by @xmath23 [ evidently , the limit @xmath24 corresponds to the limit @xmath25 . thus , in this limit , the whole region in which @xmath26 is timelike is considered . ] strictly speaking , we also have to introduce outer boundaries , say at @xmath27 ( @xmath28 ) , to control the infinite volume of the constant-@xmath16 surface . however , even if there are outer boundaries , the following arguments still hold . [ fig : kruskal ] in this situation , there is a natural choice for division of the system of the scalar field : let @xmath29 be the space of mode functions with supports in the region @xmath19 and @xmath30 be the space of mode functions with supports in the region @xmath20 . thence , the space @xmath3 of all states are of the form ( [ eqn : f = f1*f2 ] ) , where @xmath31 and @xmath32 are defined as symmetric fock spaces constructed from @xmath29 and @xmath30 , respectively : @xmath33 here @xmath34 denotes the symmetrization . let us investigate what kind of condition should be imposed for our arguments to be self - consistent . a clear condition is that the backreaction of the scalar field to the background geometry should be finite . for the brick wall model this condition is satisfied . namely , in ref . , it was shown that the total mass of the thermal atmosphere of quantum fields is actually bounded . thus , also for our system , we would like to impose the condition that the contribution @xmath35 of the subsystem @xmath31 to the mass of the background geometry should be bounded in the limit @xmath24 . it is easily shown that @xmath35 is given by @xmath36 where @xmath37 is the hamiltonian of the subsystem @xmath31 with respect to the killing time @xmath16 . hence , the expectation value of @xmath35 with respect to a state @xmath2 of the scalar field is decomposed into the contribution of excitations and the contribution from the zero - point energy : @xmath38 where @xmath39 is entanglement energy defined by @xmath40 and @xmath41 is the zero - point energy of the boulware state . here , the colons denote the usual normal ordering . [ this definition of entanglement energy corresponds to @xmath42 in ref . @xcite and @xmath43 in ref . @xcite . ] since the boulware energy @xmath41 diverges as @xmath44 in the limit @xmath24 , we should impose the condition @xmath45 where @xmath46 is the area of the horizon , @xmath47 is the hawking temperature . we would like to call this condition _ the small backreaction condition ( sbc)_. note that the right hand side of sbc ( [ eqn : sbc ] ) is independent of the state @xmath2 . now , we shall show that the hartle - hawking state is a maximum of the entanglement entropy in the space of quantum states satisfying sbc . for this purpose , we prove a more general statement for a quantum system with a state - space of the form ( [ eqn : f = f1*f2 ] ) : _ a state of the form @xmath48 is a maximum of the entanglement entropy in the space of states with fixed expectation value of the operator @xmath49 defined by @xmath50 provided that the real constant @xmath51 is determined so that the expectation value of @xmath52 is actually the fixed value . here , @xmath53 and @xmath54 ( @xmath55 ) are bases of the subspaces @xmath31 and @xmath32 , respectively , and @xmath56 are assumed to be real and non - negative . _ note that this statement is almost the same as the following statement in statistical mechanics : a canonical state is a maximum of statistical entropy in the space of states with fixed energy , provided that the temperature of the canonical state is determined so that the energy is actually the fixed value . note that the expectation value of @xmath52 is equal to the entanglement energy ( [ eqn : eent ] ) , providing that @xmath57 and @xmath56 are an eigenstate and an eigenvalue of the normal - ordered hamiltonian @xmath58 of the subsystem @xmath31 . hence , for the system of the scalar field , the above general statement insists that the state ( [ eqn : hh - state ] ) is a maximum of the entanglement entropy in the space of states satisfying sbc , which corresponds to fixing the entanglement entropy . off course , in this case , the constant @xmath51 should be determined so that sbc ( [ eqn : sbc ] ) is satisfied . returning to the subject , let us prove the general statement . ( the following proof is the almost same as that given in the appendix of ref . @xcite for a slightly different statement . however , for completeness , we shall give the proof . ) first , we decompose an element @xmath2 of @xmath3 as @xmath59 where the coefficients @xmath60 ( @xmath61 ) are complex numbers satisfying @xmath62 and can be considered as matrix elements of a matrix @xmath63 . since @xmath64 is a non - negative hermitian matrix , it can be diagonalized as @xmath65 where @xmath66 is a diagonal matrix with diagonal elements @xmath67 ( @xmath68 ) and @xmath69 is a unitary matrix . for this decomposition and diagonalization , the entanglement entropy and the expectation value of the operator @xmath52 are written as follows . @xmath70 where @xmath71 is matrix elements of @xmath69 . the constraints @xmath62 and @xmath72 are equivalent to @xmath73 next , we shall show that these expressions are equivalent to those appearing in statistical mechanics in @xmath31 . let us consider a density operator @xmath74 on @xmath75 : @xmath76 where @xmath77 is a non - negative hermitian matrix with unit trace . by diagonalizing the matrix @xmath78 as @xmath79 we obtain the following expressions for entropy @xmath80 and an expectation value @xmath81 of the operator @xmath82 . @xmath83 where @xmath84 is the diagonal elements of @xmath85 . the constraints @xmath86 and @xmath87 are restated as @xmath88 from these and those expressions , the following correspondence is easily seen : @xmath89 hence , a maximum of @xmath80 in the space of statistical states with a fixed value of @xmath81 gives a set of maxima of @xmath5 in the space of quantum states with a fixed value of @xmath39 . ( all of them are related by unitary transformations in the subspace @xmath32 . ) thus , since the thermal state @xmath90 is a maximum of @xmath80 in the space of statistical states with a fixed value of @xmath81 , @xmath91 is a maximum of @xmath5 in the space of quantum states with a fixed value of @xmath39 . here the temperature ( or the constant ) @xmath51 should be determined so that @xmath81 ( or @xmath39 ) has the fixed value . this completes the proof of the general statement . therefore , for the system of the scalar field , a state of the form ( [ eqn : hh - state ] ) is a maximum of the entanglement entropy in the space of quantum states satisfying sbc , provided that the constant @xmath51 is determined so that sbc is satisfied . the value of @xmath51 is easily determined as @xmath92 by using the well - known fact that the negative divergence in the boulware energy density can be canceled by thermal excitations if and only if temperature with respect to the time @xmath16 is equal to the hawking temperature . finally , we obtain the statement that the hartle - hawking state is a maximum of entanglement entropy in the space of quantum states satisfying sbc since the hartle - hawking state is actually of the form ( [ eqn : hh - state ] ) with @xmath92 @xcite . [ strictly speaking , in order to obtain the hartle - hawking state , we have to take the limit @xmath24 ( and @xmath93 ) . however , the following arguments still hold for a finite value of @xmath17 ( and @xmath94 ) . ] the corresponding reduced density matrix is the thermal state with temperature equal to the hawking temperature . therefore , the maximal entanglement entropy is equal to the thermal entropy with the hawking temperature , which is sought in the brick wall model . in summary the brick wall model seeks the maximal value of entanglement entropy . in other words , the entanglement approach reduces to the brick wall model when we seek the maximal entanglement entropy . our arguments suggests strong connection among three kinds of thermodynamics : black hole thermodynamics , statistical mechanics , and entanglement thermodynamics @xcite . it will be interesting to investigate close relations among them in detail . j. d. bekenstein , phys . rev . * d7 * , 949 ( 19973 ) . s. w. hawking , commun . phys . * 43 * , 199 ( 1975 ) . l. bombelli , r. k. koul , j. lee and r. d. sorkin , phys . rev . * d34 * , 373 ( 1986 ) . m. srednicki , phys . lett . * 71 * , 666 ( 1993 ) . g. thooft , nucl . phys . * b256 * , 727 ( 1985 ) . s. mukohyama and w. israel , phys . rev . * d58 * , 104005 ( 1998 ) . s. mukohyama , m. seriu and h. kodama , phys . rev . * d58 * , 064001 ( 1998 ) . s. mukohyama , _ `` the origin of black hole entropy '' _ , doctoral thesis , gr - qc/9812079 . s. mukohyama , phys . rev . * d58 * , 104023 ( 1998 ) . j. b. hartle and s. w. hawking , phys . * d13 * , 2188 ( 1976 ) . w. israel , phys . 57a * , 107 ( 1976 ) . s. mukohyama , m. seriu and h. kodama , phys . rev . * d55 * , 7666 ( 1997 ) . s. mukohyama , _ hartle - hawking state is a maximum of entanglement entropy _ , gr - qc/9904005 .
there have been many attempts to understand the statistical origin of black - hole entropy . among them , entanglement entropy and the brick wall model are strong candidates . in this paper we show a relation between entanglement entropy and the brick wall model : the brick wall model seeks the maximal value of the entanglement entropy . in other words , the entanglement approach reduces to the brick wall model when we seek the maximal entanglement entropy .
determining the conditions for epidemic extinction is an important public health problem . global eradication of an infectious disease has rarely been achieved , but it continues to be a public health goal for polio @xcite and many other diseases , including childhood diseases . more commonly , disease extinction , or fade out , is local and may be followed by a reintroduction of the disease from other regions @xcite . extinction may also occur for individual strains of a multistrain disease @xcite , such as influenza or dengue fever . since extinction occurs in finite populations , it depends critically on local community size @xcite . moreover , it is important to know how parameters affect the chance of extinction for predicting the dynamics of outbreaks and for developing control strategies to promote epidemic extinction @xcite . the determination of extinction risk is also of interest in related fields , such as evolution and ecology . for example , in the neutral theory of ecology , bio - diversity arises from the interplay between the introduction and extinction of species @xcite . in general , extinction occurs in discrete , finite populations undergoing stochastic effects due to random transitions or perturbations . the origins of stochasticity may be internal to the system or may arise from the external environment . small population size , low contact frequency for frequency - dependent transmission , competition for resources , and evolutionary pressure @xcite , as well as heterogeneity in populations and transmission @xcite , may all be determining factors for extinction to occur . extinction risk is affected by the nature and strength of the noise @xcite , as well as other factors , including outbreak amplitude @xcite and seasonal phase occurrence @xcite . for large populations , the intensity of internal population noise is generally small . however , a rare , large fluctuation can occur with non - zero probability and the system may be able to reach the extinct state . for respiratory diseases such as sars , super - spreading may account for these large stochastic fluctuations @xcite . since the extinct state is absorbing due to effective stochastic forces , eventual extinction is guaranteed when there is no source of reintroduction @xcite . however , because fade outs are usually rare events in large populations , typical time scales for extinction may be extremely long . stochastic population models of finite populations , which include extinction processes , are effectively described using the master equation formalism . stochastic master equations are commonly used in statistical physics when dealing with chemical reaction processes @xcite and predict probabilities of rare events occurring at certain times . for many problems involving extinction in large populations , if the probability distribution of the population is approximately stationary , the probability of extinction is a function that decreases exponentially with increasing population size . the exponent in this function scales as a deterministic quantity called the action @xcite . it can be shown that a trajectory that brings the system to extinction is very likely to lie along a most probable path , called the optimal extinction trajectory or optimal path . it is a remarkable property that a deterministic quantity such as the action can predict the probability of extinction , which is inherently a stochastic process @xcite . locating the optimal path is desirable because the quantity of interest , the extinction rate , depends on the probability to traverse this path , and the effect of a control strategy on extinction rate can be determined by its effect on the optimal path @xcite . by employing an optimal path formalism , we convert the stochastic problem to a mechanistic dynamical systems problem . in contrast to approaches based on diffusive processes that are valid only in the limit of large system sizes @xcite , this dynamical systems approach can give accurate estimates for the extinction time even for small populations if the action is sufficiently large . additionally , unlike other methods that are used to estimate lifetimes , this approach enables one both to estimate lifetimes and to draw conclusions about the path taken to extinction . this more detailed understanding of how extinction occurs may lead to new stochastic control strategies @xcite . in this article , we show that locating the optimal extinction trajectory can be achieved naturally by evolving a dynamical system that converges to the optimal path . the method is based on computing finite - time lyapunov exponents ( ftle ) , which have previously been used to find coherent structures in fluid flows @xcite . the ftle provides a measure of how sensitively the system s future behavior depends on its current state . we argue that the system displays maximum sensitivity near the optimal extinction trajectory , which enables us to dynamically evolve toward the optimal escape trajectory using ftle calculations . for several models of epidemics that contain internal or external noise sources , we illustrate the power of our method to locate optimal extinction trajectories . although our examples are taken from infectious disease models , the approach is general and is applicable to any extinction process or escape process . to introduce our idea of dynamically constructing an optimal path to extinction in stochastic systems , we show its application to a stochastic susceptible - infectious - recovered ( sir ) epidemiological model . details of the sir model can be found in sec . 1 of the electronic online supplement . figure [ fig : sir ] shows the probability density of extinction prehistory in the si plane . the probability density was numerically computed using @xmath0 stochastic sir trajectories that ended in extinction . trajectories are aligned at their extinction point . from the extinction point , the prehistory of each trajectory up to the last outbreak of infection is considered . small fluctuations in the infectious population are not considered in identifying the last outbreak . in this way , we restrict the analysis to the interval between the last large outbreak of infection and the extinction point . the resulting ( s , i ) pairs of susceptible and infectious individuals are then binned and plotted in the si plane @xcite . the resulting discrete density has been color coded so that the brighter regions correspond to higher density of trajectories . the figure shows that , among all the paths that the stochastic system can take to reach the extinct state , there is one path that has the highest probability of occurring . this is the optimal path to extinction . one can see that the optimal path to extinction lies on the peak of the probability density of the extinction prehistory . it should be noted that extinction for the stochastic sir model has been studied previously @xcite . the optimal path can be obtained using methods of statistical physics . in figure [ fig : sir ] , the numerical prediction of the entire optimal trajectory for the stochastic sir system has been overlaid on the probability density of extinction prehistory that was found using stochastic simulation . the trajectory spirals away from the endemic state , with larger and larger oscillations until it hits the extinct state . the agreement between the stochastically simulated optimal path to extinction and the predicted optimal path is excellent . the curve of figure [ fig : sir ] is obtained by recasting the stochastic problem in a deterministic form . the evolution of the probability of finding a stochastic system in a given state @xmath1 at a given time @xmath2 is described by the master equation @xcite . solving the master equation would provide a complete description of the time evolution of the stochastic system , but in general it is a difficult task to obtain explicit solutions for the master equation . thus , one generally resorts to approximations to the solution ; i.e. , one considers an ansatz for the probability density . in this case , since extinction of finite populations is a rare event , we will be interested in examining events that occur in the tail of the probability distribution . therefore , the distribution is assumed to take the form @xmath3 where @xmath4 is the probability density of finding the system in the state @xmath1 at time @xmath2 , @xmath5 is the size of the population , @xmath6 is the normalized state ( e.g. , in an epidemic model , the fraction of the population in the various compartments ) , and @xmath7 is a deterministic state function known in classical physics as the action . equation ( [ e : action ] ) describes the relationship between the action and the probability density and is based on an assumption for how the probability scales with the population size . the action is the negative of the natural log of the stationary probability distribution divided by the population size . therefore , the probability ( if we normalize the population ) is roughly given by the exponential of the action . intuitively , equation ( [ e : action ] ) expresses the assumption that the probability of occurrence of extreme events , such as extinction , lies in the tails of the probability distribution , which falls steeply away from the steady state . this approximation leads to a conserved quantity that is called the hamiltonian @xcite . from the hamiltonian , one can find a set of conservative ordinary differential equations ( odes ) that are known as hamilton s equations . these odes describe the time evolution of the system in terms of state variables @xmath8 , which are associated with the population in each compartment . for the sir example , @xmath8 is the vector @xmath9 . in addition to the state variables , the equations contain conjugate momenta variables , @xmath10 . the conjugate momenta , or noise , account for the uncertainty associated with the system being in a given state at a given time due to the stochastic interactions among the individuals of the population . these odes can be constructed from information in the master equation about the possible transitions and transition rates in the system . details can be found in [ sec : app - tlf ] . integration of the odes with the appropriate boundary conditions will then give the optimal evolution of the system under the influence of the noise . boundary conditions are chosen to be fixed points of the system . a typical case is shown schematically in figure [ fig : hyperbolic]a . deterministically , the endemic state is attracting and the extinct state repelling . however , introducing stochasticity allows the system to leave the deterministic manifold along an unstable direction of the endemic state , corresponding to nonzero noise . stochasticity leads to an additional extinct state which arises due to the general non - gaussian nature of the noise . for the extinction process of figure [ fig : sir ] , boundary conditions were the system leaving the endemic steady state and asymptotically approaching the stochastic extinct state . in general , the optimal extinction path is an unstable dynamical object , and this reflects extinction as a rare event . this has led many authors to consider how extinction rates scale with respect to a parameter close to a bifurcation point @xcite , where the dynamics are very slow . for an epidemic model this means that the reproductive rate @xmath11 should be greater than but very close to @xmath12 . however , most real diseases have @xmath11 larger than 1.5 , which translates into a faster growth rate from the extinct state . in general , in order to obtain analytic scaling results , one must obtain the odes for the optimal path either analytically ( using the classical theory of large fluctuations mentioned within this section and described in detail in [ sec : app - tlf ] ) or numerically ( using shooting methods for boundary value problems ) . this task may be impossible or extremely cumbersome , especially when the system is far from the bifurcation point . in the following section we demonstrate how to evolve naturally to the optimal path to extinction using a dynamical systems approach . we consider a velocity field which is defined over a finite time interval and is given by hamilton s equations of motion . such a velocity field , when starting from an initial condition , has a unique solution . the continuous dynamical system has quantities , known as lyapunov exponents , which are associated with the trajectory of the system in an infinite time limit , and which measure the average growth rates of the linearized dynamics about the trajectory . to find the finite - time lyapunov exponents ( ftle ) , one computes the lyapunov exponents on a restricted finite time interval . for each initial condition , the exponents provide a measure of its sensitivity to small perturbations . therefore , the ftle is a measure of the local sensitivity to initial data . details regarding the ftle can be found in sec . 2 of the electronic online supplement . the ftle measurements can be shown to exhibit `` ridges '' of local maxima . the ridges of the field indicate the location of attracting ( backward time ftle field ) and repelling ( forward time ftle field ) structures . in two - dimensional ( 2d ) space , the ridge is a curve which locally maximizes the ftle field so that transverse to the ridge one finds the ftle to be a local maximum . what is remarkable is that the ftle ridges correspond to the optimal path trajectories , which we heuristically argue in sec . 3 of the electronic online supplement . the basic idea is that since the optimal path is inherently unstable , the ftle shows that , locally , the path is also the most sensitive to initial data . figure [ fig : hyperbolic]b shows a schematic that demonstrates why the optimal path has a local maximum to sensitivity . if one chooses an initial point on either side of the path near the endemic state , the two trajectories will separate exponentially in time . this is due to the fact that both extinct and endemic states are unstable , and the connecting trajectory defining the path is unstable as well . any initial points starting near the optimal path will leave the neighborhood in short time . we now apply our theory of dynamical sensitivity to the problem of locating optimal paths to extinction for several examples . we consider the case of internal fluctuations , where the noise is not known a priori , as well as the case of external noise . in each case , the interaction of the noise and state of the system begins by finding the equations of motion that describe the unstable flow . these equations of motion are then used to compute the ridges corresponding to maximum ftle , which in turn correspond to the optimal extinction paths @xcite . for an example of a system with internal fluctuations which has an analytical solution , consider extinction in the stochastic branching - annihilation process @xmath13 where @xmath14 and @xmath15 are constant reaction rates @xcite . equation ( [ e : bd ] ) is a single species birth - death process and can be thought of as a simplified form of the verhulst logistic model for population growth @xcite . the mean field equation for the average number of individuals @xmath16 in the infinite population limit is given by @xmath17 . the stochastic process given by equation ( [ e : bd ] ) contains intrinsic noise which arises from the randomness of the reactions and the fact that the population consists of discrete individuals . this intrinsic noise can generate a rare sequence of events that causes the system to evolve to the extinct state . the probability @xmath18 to observe , at time @xmath2 , @xmath16 individuals is governed by the master equation @xmath19+\lambda\left[(n-1)p_{n-1}-np_{n}\right].\label{e : bd_master}\ ] ] the hamiltonian associated with this system is @xmath20 where @xmath21 is a conjugate coordinate related to @xmath16 through a transformation @xcite , and @xmath22 plays the role of the momentum . the equations of motion are given by = & = q[(1 + 2p)-(1+p)q],[e : bd_qdot ] + = & -=p[(2+p)q-(1+p)].[e : bd_pdot ] the mean field is retrieved in equation ( [ e : bd_qdot ] ) when @xmath23 ( no fluctuations or noise ) . the hamiltonian has three zero - energy curves . the first is the mean - field zero - energy line @xmath23 ( no fluctuations ) , which contains two unstable points @xmath24 and @xmath25 . the second is the extinction line @xmath26 ( trivial solution ) , which contains another unstable point @xmath27 . the third zero - energy curve is non - trivial and is @xmath28 the segment of the curve given by equation ( [ e : bd_op ] ) which lies between @xmath29 corresponds to a ( heteroclinic ) trajectory which exits , at @xmath30 , the point @xmath31 along its unstable manifold and enters , at @xmath32 , the point @xmath33 along its stable manifold . this trajectory is the optimal path to extinction and describes the most probable sequence of events which evolves the system from a quasi - stationary state to extinction @xcite . to show that the ftle evolves to the optimal path , we calculate the ftle field using the system of hamilton s equations given by equations ( [ e : bd_qdot])-([e : bd_pdot ] ) . figure [ fig : ftle_bd_sis]a shows both the forward and backward ftle plot computed for the finite time @xmath34 , with @xmath35 and @xmath36 . in this example , as well as the following two examples , @xmath37 was chosen to be sufficiently large so that one obtains a measurable exponential separation of trajectories . in figure [ fig : ftle_bd_sis]a , one can see that the optimal path to extinction is given by the ridge associated with the maximum ftle . in fact , by overlaying the forward and backward ftle fields , one can see all three zero - energy curves including the optimal path to extinction . also shown in figure [ fig : ftle_bd_sis]a are the analytical solutions to the three zero - energy curves given by @xmath23 , @xmath26 , and equation ( [ e : bd_op ] ) . there is excellent agreement between the analytical solutions of all three curves and the ridges which are found through numerical computation of the ftle flow fields . it is possible to compute analytically the action along the optimal path for a range of @xmath38 values . using equation ( [ e : bd_op ] ) , it is easy to show that the action @xmath7 is @xmath39 it is clear from equation ( [ e : bd_action ] ) that the action scales linearly with @xmath38 . we now consider the well - known problem of extinction in a susceptible - infectious - susceptible ( sis ) epidemiological model , which is a core model for the basis of many recurrent epidemics . the sis model is described by the following system of equations : = & -s + i - is,[e : siex_a ] + = & -(+)i + is,[e : siex_b ] where @xmath40 denotes a constant birth and death rate , @xmath41 represents the contact rate , and @xmath42 denotes the recovery rate . assuming the total population size is constant and can be normalized to @xmath43 , then equations ( [ e : siex_a])-([e : siex_b ] ) can be rewritten as the the following one - dimensional ( 1d ) process for the fraction of infectious individuals in the population : @xmath44 the stochastic version of equation ( [ e:1dsis_det ] ) is given as @xmath45 where @xmath46 is uncorrelated gaussian noise with zero mean and @xmath47 is the standard deviation of the noise intensity . the noise models random migration to and from another population @xcite . equation ( [ e:1dsis_stoch ] ) has two equilibrium points given by @xmath48 ( disease - free state ) and @xmath49 ( endemic state ) . using the euler - lagrange equation of motion @xcite from the lagrangian determined by equation ( [ e:1dsis_stoch ] ) ( @xmath50^{2}=[\dot{i}-f(i)]^{2}$ ] ) along with the boundary conditions given by the extinct and endemic states , one finds that the optimal path to extinction ( as well as its counterpart path from the disease - free state to the endemic state ) is given by @xmath51 . as in the first example , one can numerically compute the optimal path to extinction using the ftle . figure [ fig : ftle_bd_sis]b shows the forward and backward ftle plot computed for @xmath52 , with @xmath53 and @xmath54 . note that we can consider the combination @xmath55 since the lagrangian depends only on the combination rather than on either parameter separately . in figure [ fig : ftle_bd_sis]b , one can see that the optimal path from the endemic state to the disease - free state is given by the ridge associated with the local maximum ftle . also shown in figure [ fig : ftle_bd_sis]b is the counterpart optimal path from the disease - free state to the endemic state ( found by computing the backward ftle field ) . in addition , the agreement with the analytical prediction is excellent , as shown in figure [ fig : ftle_bd_sis]b . if one solves equation ( [ e:1dsis_stoch ] ) for @xmath56 and substitutes both @xmath56 and @xmath57 into the lagrangian , then one can analytically find an expression for the action along the optimal path . the expression for the action is a function of @xmath58 and the reproductive number @xmath11 and is given by @xmath59 where @xmath60 . we next consider the 1d stochastic version of the sis epidemic model for a finite population with internal fluctuations using the transition rates found in sec . 4 of the electronic online supplement . using the formalism of @xcite , one then has the following hamiltonian associated with this model : @xmath61 where @xmath62 is the fraction of infectious individuals and @xmath22 is the momentum . internal fluctuations arise from the random interactions of the population . although there is no analytical solution for the optimal path to extinction , we can once again determine the optimal path by computing the ftle flow field associated with this system . figure [ fig : ftle_ham ] shows the forward ftle plot computed using hamilton s equation of motion for @xmath63 , with @xmath64 and @xmath65 , and as in previous examples , the optimal path to extinction from the endemic state to the disease - free state is apparent . note that the non - zero momentum corresponding to the extinct state qualitatively agrees with similar boundary conditions found in @xcite and is associated with non - zero probability flux . once again , it is possible to compute the action along the optimal path for a range of values of the reproductive number @xmath11 . in contrast to the prior two examples , here the action must be computed numerically . moreover , even the optimal path must be found numerically using the ftle plot generated for each value of @xmath11 . given an @xmath11 , we computed the forward ftle flow field using a grid resolution of @xmath66 in both position and momentum space . to find the optimal path corresponding to the ridge of maximal ftle values , we let the deterministic , endemic steady state be the starting point @xmath67 of the path . then a second point @xmath68 on the path was found by locating the maximum of the ftle values in an arc of radius @xmath69 grid points spanning nearly @xmath70 radians for negative momentum values and originating at @xmath67 . subsequent points @xmath71 were found by taking the maximum ftle value on an arc of radius @xmath72 grid points spanning nearly @xmath70 radians originating at the most recent point @xmath73 and centered around the vector @xmath74 until the extinction line was reached . the complete optimal path was estimated from the @xmath73 using cubic spline interpolation . once the optimal path had been found , the action was computed by numerically integrating the lagrangian along this path . figure [ fig : action_v_r0]a shows the numerically computed action versus reproductive number over the range @xmath75 . the inset of figure [ fig : action_v_r0]a shows the portion of figure [ fig : action_v_r0]a from @xmath76 . also shown in the inset of figure [ fig : action_v_r0]a is an analytical , asymptotic scaling result that is valid for values of @xmath11 close to unity . as can be see in figure [ fig : action_v_r0]a , there is good agreement between the numerically computed action and the analytically computed action near @xmath77 . details of the derivation of the analytical scaling law can be found in sec . 5 of the electronic online supplement . figure [ fig : action_v_r0]b shows the numerically simulated mean extinction time versus reproductive number for the 1d stochastic sis model for a finite population ( see sec . [ sec : apps][sec : sis_h ] ) . also shown in figure [ fig : action_v_r0]b is the analytical prediction found using the previously mentioned scaling law derived for values of @xmath11 close to unity . as one can see , the agreement is excellent . in all of the examples of sec . [ sec : apps ] , we have shown that the optimal path to extinction is equated with having a ( locally ) maximal sensitivity to initial conditions . even though there exist many possible paths to extinction , the dynamical systems approach converges to the path that maximizes extinction . the parameter values chosen for the three examples are such that the extinct and endemic states are far away from each other . therefore , in general , there will be no possible approximate analytical treatment as performed in @xcite . in addition , we have shown how to constructively compute numerically the action for a wide range of reproductive numbers . our method allows for the computation of extinction times and can be extended to high - dimensional problems . because the method is general , and unifies dynamical systems theory with the probability of extinction , we expect that any system found in other fields can be understood using this approach . indeed , in problems of general extinction , it is now possible to evolve naturally to the optimal path , and thus predict the path that maximizes the probability to extinction . future work in this area will include improved sampling methods to find the optimal path more efficiently in higher dimensional models . specific applications of optimal path location in the future will include spatio - temporal extinction processes such as those that occur in pre - vaccine measles @xcite and multi - strain extinction in diseases such as dengue @xcite . finally , the optimal path method may lead to novel monitoring and control strategies . in one biological application , knowledge of the most probable path to extinction may help with monitoring an environment and with providing an estimate of the likelihood of extinction based on where the population lies relative to the path . it is known that once a trajectory has left the neighborhood of the endemic state , most paths to extinction occur near the optimal path . this phenomenon can be seen in figure [ fig : sir ] , where the optimal path lies on the peak of the probability density of extinction prehistory . therefore , the optimal path provides a good location to monitor the system for possible extinction events . furthermore , although the momenta ( noise effects ) are not directly observable , it may be possible to infer them dynamically @xcite from time series data of observable quantities in conjunction with the equations for the time evolution of position and momentum variables . knowledge of where a system lies in position - momentum space could provide an estimate for how quickly a desired epidemic extinction could occur or could provide the risk of extinction for a species one wishes to conserve . in yet another application , knowledge of the optimal path to extinction has the advantage that , in the presence of noise ( that is estimated from data ) and a known population of infectious individuals , it may be possible to develop better vaccine controls that reduce the time to extinction . figure [ fig : hyperbolic](a ) shows a schematic of the path to extinction , where the extinct state is a saddle with stable and unstable directions . for many epidemic models , the extinct state can be shown to have a similar geometry of stable and unstable directions . an approach to the extinct state on the optimal path will lead to the fastest time to extinction . moreover , since the extinct state has a saddle structure in the presence of noise , it may be controlled with projection methods @xcite or probabilistic techniques @xcite . one can consider instituting a method of parameter control , where the parameters could be vaccine levels or treatment of infectious individuals . combined with the above - mentioned monitoring techniques , the control method will allow one to move an existing state that deviates from the optimal path closer to the optimal path as time evolves . by adjusting the parameters , we may target the stable directions of the path when we are close to epidemic extinction @xcite . comparing observations with model predictions of the optimal path allow us to use controls to minimize the time to extinction . the authors gratefully acknowledge support from the office of naval research , the air force office of scientific research , the army research office , and the national institutes of health . i.b.s and l.b.s . are supported by award number r01gm090204 from the national institute of general medical sciences . the content is solely the responsibility of the authors and does not necessarily represent the official views of the national institute of general medical sciences or the national institutes of health . we also gratefully acknowledge m. dykman for helpful discussions . [ sec : app - tlf ] letting @xmath5 denote the population size , the state variables @xmath78 of the system describe the components of a population , while the random state transitions which govern the dynamics are described by the transition rates @xmath79 , where @xmath80 is an increment in the change of @xmath81 . in the large population limit without any fluctuations , the mean field equations are given by the system @xmath82 consideration of the net change in increments of the state of the system results in the following master equation for the probability density @xmath83 of finding the system in state @xmath81 at time @xmath2 : @xmath84.\label{e : masterequation}\\\ ] ] the solution to the master equation has a peak around a stable steady state in the limit of large population ( @xmath85 ) with width @xmath86 @xcite . however , since we are interested in the probability of extinction , we will consider the tail of the distribution , which gives the probability of having a small number of individuals . the tail of the distribution can be obtained by employing the ansatz given by equation , which is an assumption for how the probability density scales with population size @xcite . equation also implies that a maximum in the extinction probability can be found minimizing the action over a set of extinction paths starting from the stable endemic state . the assumption of this functional form for @xmath87 allows the action to be derived from properties of the master equation . the density , @xmath83 , can be found by substituting the ansatz given by equation into the master equation . the resulting equations for the action are given by the hamilton - jacobi equation for a hamiltonian , @xmath88 , given by @xmath89 where we have normalized the population components and rates respectively as @xmath90 . following the classical mechanics convention , define a conjugate momentum to the state space , @xmath91 , by letting @xmath92 and where @xmath93 is the classical hamiltonian @xcite . the hamiltonian , @xmath88 , depends both on the state of the system , @xmath94 , as well as the momentum , @xmath95 , which provides an effective force due to stochastic fluctuations on the system . the hamiltonian equations of motion provide trajectories in time of @xmath96 and @xmath97 , and as such , describe a set of paths going from an initial point at time @xmath98 to some final point at time @xmath99 in @xmath100 space . for a given path , the action is given by @xmath101 , and as such determines the exponent of the probability of observing that path . ( it should be noted that instead of using the hamiltonian representation , one could use the lagrangian representation @xmath102 , which results in an equivalent solution . ) the action reveals much information about the probability evolution of the system , from scaling near bifurcation points in non - gaussian processes to rates of extinction as a function of epidemiological parameters @xcite . as already stated , in order to maximize the probability of extinction , one must minimize the action @xmath7 . the minimizing formulation entails finding the solution to the hamilton - jacobi equation , which means that one must solve the @xmath103-dimensional system of hamilton s equations [ @xmath104 , @xmath105 for @xmath91 and @xmath106 , where the hamiltonian is explicitly given as @xmath107.\label{e : hamilonian}\ ] ] the appropriate boundary conditions of the system are such that a solution starts at a non - zero state , such as an endemic state , and asymptotically approaches one or more zero components of the state vector , representing a disease free state . therefore , a trajectory that is a solution to the two - point boundary value problem determines a path , which in turn yields the probability of going from the initial state to the final state . the optimal path to extinction is the path which minimizes the action in either the hamiltonian or lagrangian representation . we compute the trajectory satisfying the hamiltonian system that has as its asymptotic limits in time the endemic state as @xmath108 and the extinct state as @xmath109 . the momentum @xmath110 represents the force of the fluctuations on the population , and this momentum changes the stability of the equilibrium points . both the endemic and extinct states have attracting and repelling directions for @xmath111 , as shown schematically in figure [ fig : hyperbolic ] . allen , l. j. s. & burgin , a. m. 2000 comparison of deterministic and stochastic sis and sir models in discrete time . * 163 * , 133 . alonso , d. , mckane , a. j. & pascual , m. 2006 stochastic amplification in epidemics . * 4 * , 575582 . doering , c. r. , sargsyan , k. v. & sander , l. m. 2005 extinction times for birth - death processes : exact results , continuum asymptotics , and the failure of the fokker - planck approximation . * 3*(2 ) , 283299 . schwartz , i. b. & triandaf , i. 1994 control and prediction in seasonally driven population models . in _ predictability and nonlinear modelling in natural sciences and economics _ grasman , j. and van straten , g. ) , pp . kluwer academic . stochastic realizations . the results were computed using monte carlo simulations , and details of the sampling of the trajectories are described in the text . the curve is the numerically predicted optimal path to extinction . note that the optimal path to extinction lies on the peak of the probability density of extinction prehistory . the population is @xmath114 individuals , with @xmath115 ( contact rate @xmath116 , recovery rate @xmath117 , birth - death rate @xmath118 ) . , width=321,height=188 ] ) coordinate . the deterministic manifold ( @xmath23 ) is indicated in blue . deterministically , the extinct state is repelling and endemic state is attracting . however , the endemic state has unstable directions for nonzero noise ( @xmath119 ) , and the optimal path ( red ) is the trajectory carrying the system from the endemic state to a stochastic extinct state . b ) schematic showing the path from the endemic state ( blue ) to the extinct state ( red ) . the optimal path leaves the endemic point along an unstable manifold , and connects to the extinct state along its stable manifold . the magenta and green dashed lines represent trajectories initially separated by the optimal path . the initial starting distance between trajectories near the endemic state expands exponentially in forward time ( shown by the cyan lines ) . locally , this shows that the finite - time lyapunov measure of sensitivity with respect to initial data is maximal along the optimal path.,title="fig:",width=226 ] ) coordinate . the deterministic manifold ( @xmath23 ) is indicated in blue . deterministically , the extinct state is repelling and endemic state is attracting . however , the endemic state has unstable directions for nonzero noise ( @xmath119 ) , and the optimal path ( red ) is the trajectory carrying the system from the endemic state to a stochastic extinct state . b ) schematic showing the path from the endemic state ( blue ) to the extinct state ( red ) . the optimal path leaves the endemic point along an unstable manifold , and connects to the extinct state along its stable manifold . the magenta and green dashed lines represent trajectories initially separated by the optimal path . the initial starting distance between trajectories near the endemic state expands exponentially in forward time ( shown by the cyan lines ) . locally , this shows that the finite - time lyapunov measure of sensitivity with respect to initial data is maximal along the optimal path.,title="fig:",width=226 ] ) with @xmath35 and @xmath36 . the integration time is @xmath34 with an integration step size of @xmath120 and a grid resolution of @xmath121 in both @xmath21 and @xmath22 ( momentum ) . the three zero - energy curves are given by the ridges of maximal ftle and are overlaid with the analytical solution of these curves given by @xmath23 , @xmath26 , and equation ( [ e : bd_op ] ) . ( b ) forward and backward ftle flow fields for the sis epidemic model with external fluctuations . the flow fields were computed using the lagrangian given by equation ( [ e:1dsis_stoch ] ) with @xmath53 and @xmath54 . the integration time is @xmath52 with an integration step size of @xmath120 and a grid resolution of @xmath121 in both @xmath62 and @xmath22 ( momentum ) . the optimal path to extinction from the endemic state to the disease - free state and its counterpart optimal path from the disease - free state to the endemic state ( found by computing the backward ftle field ) are given by the ridges of maximal ftle and are overlaid with the analytical solution of the optimal paths.,width=236 ] ) with @xmath35 and @xmath36 . the integration time is @xmath34 with an integration step size of @xmath120 and a grid resolution of @xmath121 in both @xmath21 and @xmath22 ( momentum ) . the three zero - energy curves are given by the ridges of maximal ftle and are overlaid with the analytical solution of these curves given by @xmath23 , @xmath26 , and equation ( [ e : bd_op ] ) . ( b ) forward and backward ftle flow fields for the sis epidemic model with external fluctuations . the flow fields were computed using the lagrangian given by equation ( [ e:1dsis_stoch ] ) with @xmath53 and @xmath54 . the integration time is @xmath52 with an integration step size of @xmath120 and a grid resolution of @xmath121 in both @xmath62 and @xmath22 ( momentum ) . the optimal path to extinction from the endemic state to the disease - free state and its counterpart optimal path from the disease - free state to the endemic state ( found by computing the backward ftle field ) are given by the ridges of maximal ftle and are overlaid with the analytical solution of the optimal paths.,width=236 ] with @xmath64 and @xmath54 . the integration time is @xmath63 with an integration step size of @xmath120 and a grid resolution of @xmath66 in both @xmath62 and @xmath22 ( momentum ) . the optimal path to extinction is given by the ridge of maximal ftle.,width=321 ] ( a ) numerically computed action ( integrated along the optimal path found numerically from the ftle flow field ) versus reproductive number @xmath11 for the sis epidemic model with internal fluctuations . the inset shows a portion of the graph near @xmath77 . the numerically computed action is given by the black points , while the dashed , red curve shows an asymptotic scaling result that is valid near @xmath77 , and is given by @xmath122 $ ] ( see sec . 5 of the electronic online supplement ) . ( b ) numerically simulated ( solid curve with black points ) mean extinction time versus reproductive number for the sis epidemic model with internal noise and the analytical prediction ( dashed , red curve ) found using the asymptotic scaling law that is valid near @xmath77.,width=236 ] ( a ) numerically computed action ( integrated along the optimal path found numerically from the ftle flow field ) versus reproductive number @xmath11 for the sis epidemic model with internal fluctuations . the inset shows a portion of the graph near @xmath77 . the numerically computed action is given by the black points , while the dashed , red curve shows an asymptotic scaling result that is valid near @xmath77 , and is given by @xmath122 $ ] ( see sec . 5 of the electronic online supplement ) . ( b ) numerically simulated ( solid curve with black points ) mean extinction time versus reproductive number for the sis epidemic model with internal noise and the analytical prediction ( dashed , red curve ) found using the asymptotic scaling law that is valid near @xmath77.,width=236,height=158 ]
extinction , optimal path , finite - time lyapunov exponents extinction appears ubiquitously in many fields , including chemical reactions , population biology , evolution , and epidemiology . even though extinction as a random process is a rare event , its occurrence is observed in large finite populations . extinction occurs when fluctuations due to random transitions act as an effective force which drives one or more components or species to vanish . although there are many random paths to an extinct state , there is an optimal path that maximizes the probability to extinction . in this article , we show that the optimal path is associated with the dynamical systems idea of having maximum sensitive dependence to initial conditions . using the equivalence between the sensitive dependence and the path to extinction , we show that the dynamical systems picture of extinction evolves naturally toward the optimal path in several stochastic models of epidemics . [ firstpage ]
thermal gauge field theories in the temporal gauge has been the subject of many investigations both in the imaginary and in the real time formalisms @xcite . one of the main advantages of the _ non - covariant _ temporal gauge is that it is physical and effectively ghost - free . at finite temperature , it may be considered a more natural choice , since the lorentz invariance is already broken by the presence of the heat bath . it is also convenient for calculating the response of the qcd plasma to a chromo - electric field , since in this case only the gluon self - energy is needed ( in this gauge , the chromo - electric field depends only linearly on the gauge field @xmath3 ) @xcite . despite these advantages , explicit calculations of loop corrections to green functions are known to be more complicated than in covariant gauges . one of the difficulty is associated with the more involved tensorial structure of the propagator . the other more fundamental problem is how to deal with the extra poles at @xmath4 in the gluon propagator , where @xmath5 is the loop momentum and @xmath6 ( @xmath7 ) is the _ temporal axial _ four - vector . in the imaginary time formalism , the standard method of calculation of thermal green functions consists in replacing the matsubara frequency sum by a _ contour integral _ in the complex plane @xcite . using this approach , all green functions , at one - loop order , can be written as the sum of the _ vacuum part _ plus the a _ thermal part _ which involves the bose - einstein ( boson loop ) or fermi - dirac ( fermion loop ) statistical distributions . in the case of the temporal gauge , the extra poles at @xmath4 in the gluon propagator demand more care and make the standard _ contour integral method _ of calculation very involved . in order to deal with this problem , leibbrandt and staley @xcite have developed a technique for performing perturbative calculations at finite temperature in the temporal gauge ( see also reference @xcite for a different approach to this problem ) . employing a special version of the _ mandelstam - leibbrandt prescription _ @xcite combined with the _ @xmath8-function method _ @xcite , they have obtained the _ complete _ @xmath9-expansion for the one - loop self - energy component @xmath10 . their results show that the leading and sub - leading contributions to the self - energy can be unambiguously determined in the temporal gauge , and are in agreement with the standard _ contour integral _ method employed in reference @xcite . the temperature dependent part of green functions can be described in terms of _ forward scattering amplitudes _ of on - shell thermal particles of the thermal medium . this idea was described in reference @xcite and has been further elaborated in the _ feynman gauge _ and shown to be very useful for determining the partition function in qcd at high temperature @xcite . more recently this result has been generalized to the case of _ general covariant gauges _ @xcite . this method has been derived using both the imaginary time formalism and the real time formalism up to two - loop order @xcite . there is also an interesting relation with the path integral approach @xcite . one of the purposes of this work is to extend the forward scattering method to a class of temporal gauges . the main advantage of this method is that in the high - temperature limit , it is straightforward to obtain the _ full tensor structure _ of both the leading @xmath0 and the sub - leading logarithmic contributions for temperatures @xmath2 such that @xmath11 , where @xmath12 denotes the external four - momentum . for the leading @xmath0 part we obtain the well known gauge invariant _ hard thermal loop _ result . the _ gauge dependent _ sub - leading logarithmic part shares with the previous calculations in general covariant gauges @xcite the interesting property of having the same structure as the _ ultraviolet pole contributions _ which occur at zero temperature . another purpose of this work is to study the transversality property of the thermal gluon self - energy . as is well known @xcite , in general covariant gauges , the exact self - energy of the gluon is not transverse at finite temperature . in section ii we derive , in a class of temporal gauges , the full tensor structure of the forward scattering amplitude associated with the gluon self - energy . in section iii we show that , in contrast to the behavior in general covariant gauges , the exact thermal self - energy of the gluon is transverse to all orders . in section iv we compute the leading gauge - invariant @xmath0 and the sub - leading @xmath13 contributions of the non - vanishing transverse structures . the later is shown to be simply related to the ultraviolet pole terms which occur at zero temperature . finally , we present a brief conclusion in section v. the diagrams which contribute to the gluon self - energy are shown in the figure 1 . we employ the feynman rules for the three- and four gluon vertices given in reference @xcite . in the class of temporal gauges characterized by a gauge parameter @xmath14 , the gluon propagator can be written as @xmath15\right\ } \ , , \ ] ] where the usual axial vector has been written as @xmath16 . [ fig1 ] before considering the more involved diagrams in figures ( 1a ) and ( 1b ) , let us show that the finite temperature contribution to the diagram in figure ( 1c ) vanishes . in the imaginary time formalism this contribution is proportional to the following integral @xcite @xmath17,\ ] ] where @xmath18 is the bose - einstein statistical distribution and @xmath19 is a _ momentum independent _ tensor which is obtained from the explicit expression for gluon - ghost vertex ( note that the gluon - ghost vertex is not equal to zero in the class of gauges considered here ) . using the partial fraction decomposition ( for @xmath20 different from zero ) @xmath21\ ] ] and making the change of variable @xmath22 in the second term on the right hand side , we can see that eq . ( [ eq2 ] ) _ vanishes _ when one takes into account the property @xmath23 this result holds independently of any explicit prescription for the poles at @xmath24 . let us now consider the diagrams in figures ( 1a ) and ( 1b ) . we are specifically interested in those contributions which can be expressed in terms of forward scattering amplitudes of _ on - shell _ thermal particles . following the steps described in the appendix [ appa ] , such an expression can be obtained closing the contour of the @xmath25-integration in the right hand side of the complex plane and considering _ only _ the poles from the propagator ( [ propag ] ) at @xmath26 and at @xmath27 , respectively for the diagrams in the figures ( 1a ) and ( 1b ) . after shifting the momentum in the contribution from the second pole in the diagram of figure ( 1a ) and using the property ( [ eq5 ] ) , we obtain the following expression for the forward scattering ( fs ) part of the gluon self - energy @xmath28{c } \epsfbox{gluonforward.ps } \end{array } + q\leftrightarrow -q \right\}_{q_0=|\vec q|},\ ] ] where the factor @xmath29 in front of the curly bracket takes into account the symmetry of the graphs in figures ( 1a ) and ( 1b ) . it is understood that the lorentz and color indices of the thermal particles ( horizontal lines ) are contracted with the tensor given by the curly bracket of eq . ( [ propag ] ) . the full tensor structure generated by expression ( [ eq6 ] ) is rather lengthy . in order to express the result in a more suggestive form , it is convenient to use the following tensor decomposition @xcite @xmath30 where @xmath31 is a set of four independent tensors which can be constructed using two lorentz indices , the gluon four momentum @xmath12 and a vector @xmath32 which describes the four velocity of the heat bath . the tensors @xmath33 are transverse with respect to the external four - momentum @xmath34 and satisfy @xmath35 and @xmath36 ( @xmath37 ) , whereas the tensors @xmath38 and @xmath39 are non - transverse . using the orthogonality of the tensors given in eq . ( [ eq7 ] ) as well as the normalization @xmath40 , it is straightforward to obtain the explicit expressions for @xmath41 , @xmath42 , @xmath43 , @xmath44 from eqs . ( [ eq6 ] ) and ( [ eq7a ] ) ( the tensor algebra was performed using the maplevr3 version of the symbolic computer package hip @xcite ) . the resulting expressions for @xmath43 and @xmath44 turn out to be _ equal to zero_. the vanishing of the one - loop correction to @xmath44 is a general consequence of the finite temperature slavnov - taylor identities @xcite while the vanishing of @xmath43 is a property which is verified by the forward scattering part of the gluon self - energy . from the transversality property of @xmath45 and @xmath46 it follows that @xmath47 in order to ascertain whether the transversality property of the thermal gluon self - energy holds beyond the approximation of forward amplitudes with on - shell particles , it is useful to study the consequences of the becchi - rouet - stora identities @xcite on the structure of @xmath48 at finite temperature . to derive these , we start from the effective action : @xmath49 where the ellipsis denote ghost terms which are unimportant for our purpose @xcite . @xmath50 is the source of the brs transformation , and @xmath51 is a ghost field ( although in our case the closed ghost loops do not contribute to feynman diagrams @xcite , the ghost fields need not necessarily decouple from the brs identity , which involves open ghost lines ) . @xmath52 is a quantity related to the gauge transformation of the gluon field @xmath53 , being given to lowest order by the covariant derivative @xmath54 @xcite . the quadratic part @xmath55 is the sum of free kinetic energy , without the gauge fixing term and the self - energy : @xmath56 the relevant brs invariance of the gluon self - energy can now be written as @xmath57 differentiating this equation with respect to @xmath58 and @xmath59 , and setting the sources and fields equal to zero leads in momentum space to the identity : @xmath60 since , to lowest order , @xmath61 is proportional to @xmath62 , it follows from eq . ( [ teq2 ] ) that the above identity implies the relation : @xmath63 where @xmath64 denotes the contributions of feynman loops to @xmath65 . let us first consider the one loop order function @xmath66 , which may be represented as shown in figure 2a . [ fig2 ] using eq . ( [ propag ] ) for gluon propagator and the fact that the gluon - ghost vertex is proportional to @xmath67 , one finds that @xmath66 is given at finite temperature by @xmath68}\ ] ] where @xmath69 is the number of colors , the space - time dimension is @xmath70 and @xmath71 . one can now employ the identity ( [ eq4 ] ) and make the change of variable @xmath72 in the second term , to get @xmath73 since this expression is orthogonal to @xmath74 , it follows from eq . ( [ teq5 ] ) , that the exact gluon self - energy is transverse to one - loop order . in order to study the behavior of @xmath75 in higher orders , let us consider , for example , the two - loops diagram shown in figure 2b . this yields at finite temperature a contribution given by : @xmath76 \\ \nonumber & & \times\sum^{+\infty}_{l=-\infty}\int { d^{3-\epsilon}q \over ( 2\pi)^{3-\epsilon } } { q_{\sigma}\over ( q\cdot n)[(q+k)\cdot n ] [ ( p - q)\cdot n]}\end{aligned}\ ] ] where @xmath77 and @xmath78 . using partial fraction decompositions and making appropriate changes of variables , one finds that eq . ( [ eq.1 ] ) can be written in the form : @xmath79 \\ \nonumber & & \times \sum^{+\infty}_{l=-\infty}\int { d^{3-\epsilon}q \over ( 2\pi)^{3-\epsilon } } { 1\over ( q\cdot n)}\end{aligned}\ ] ] which is rather typical of the contributions associated with two - loop diagrams . comparing the relation ( [ eq.2 ] ) with the one - loop result given by eq . ( [ teq7 ] ) , one can see that both expressions have the same @xmath5-dependence . but such contributions vanish , in consequence of the antisymmetry of @xmath80 . note that , thus far , we have made no assumptions about the specific form of @xmath81 , so that this conclusion is also true in a class of axial gauges @xcite . in the specific case of temporal gauges where @xmath82,it is worth noticing that the integrals decouple , in fact , from the sums . then , we can set , in the sense of dimensional regularisation : @xmath83 furthermore , using the mandelstam - leibbrandt prescription to remove the singular terms at @xmath84 and @xmath85 , the sum over @xmath86 also gives zero because of the antisymmetry of @xmath78 . the above arguments , which show that @xmath66 and @xmath87 are equal to zero , may be extended to higher order to establish that @xmath88 must generally vanish . it then follows from eq . ( [ teq5 ] ) that the exact self - energy of the gluon is transverse at all temperatures . this important property implies that , to all orders , the structure functions @xmath89 and @xmath90 should be equal to zero . ( we show in the appendix [ appb ] that these results are confirmed by explicit calculations to one - loop order . ) in consequence of this behavior , the non - linear relation among the structure functions @xcite @xmath91 which must hold at finite temperature in any linear gauge , is manifestly satisfied in the class of temporal gauges . the non - vanishing structures @xmath92 and @xmath93 , can be expanded as a series of high - temperature terms which come from the _ hard thermal loop _ region of _ large _ @xmath5 , such that @xmath94 . ( there are also other contributions , such as those proportional to @xmath2 , which originate from the region of small @xmath5 @xcite . ) this can be done in a straightforward way expressing the integrands in terms of the dimensionless four - vector @xmath95 and performing the expansion in powers of @xmath96 . the results up to sub - leading contributions are given by @xmath97 \int d|\vec q|\,|\vec q|\,n(|\vec q| ) \right . \nonumber \\ & & + \left[-4q\cdot kk^{2}k_{0 } -{k^{8}\over ( q\cdot k)^{4 } } + 2{k^{6}k_{0}\over ( q\cdot k)^{3 } } + 4k^{2}k_{0}^{2 } -8{k^{4}k_{0}^{2}\over ( q\cdot k)^{2 } } + 2k^{4 } + { 5k^{6}\over ( q\cdot k)^{2 } } \right . \nonumber \\ & & \left.\left . + { \alpha\over n_{0}^{2}}\left ( 4 k^{4}k_{0}^{2 } -2 k^{6 } -4 q\cdot k k^{4}k_{0 } + 2 ( q\cdot k)^{2}k^{4 } \right ) \right ] \int{d|\vec q|\over |\vec q|}n(|\vec q| ) \right\}\end{aligned}\ ] ] and @xmath98 \int d|\vec q|\,|\vec q|\,n(|\vec q| ) \right . \nonumber \\ & & + \left[8q\cdot kk^{2}k_{0 } -{4k^{6}k_{0}\over ( q\cdot k)^{3 } } -16k^{2}k_{0}^{2 } + { 2k^{4}k_{0}^{2}\over ( q\cdot k)^{2 } } + 4k^{4 } + { 2k^{8}\over ( q\cdot k)^{4 } } + { 4k^{6}\over ( q\cdot k)^{2 } } \right . \nonumber \\ & & \left.\left . -{\alpha\over n_{0}^{2 } } \left ( 4 k^{4}k_{0}^{2 } + 4 k^{2}(q\cdot k)^{2}k_{0}^{2 } -8 q\cdot k k^{2}k_{0}^{3 } \right ) \right ] \int{d|\vec q|\over |\vec q|}n(|\vec q|)\right\ } \ , , \end{aligned}\ ] ] where @xmath99 denotes the integration over the directions of @xmath100 . using that @xmath101 and performing the angular integrations , we obtain @xmath102\right . \nonumber \\ -\left.{k^{2}\over 12\pi^{2}}\left(11 - 4{\alpha\over n_{0}^{2 } } k^{2}\right ) \int { d |\vec q|\over |\vec q|}n(|\vec q| ) \right\}\end{aligned}\ ] ] @xmath103 \right . + \left.{k^{2}\over 12\pi^{2}}\left(11 - 4{\alpha\over n_{0}^{2 } } k_{0}^{2}\right ) \int{d |\vec q|\over |\vec q|}n(|\vec q| ) \right\ } \ , .\end{aligned}\ ] ] the above contributions proportional to @xmath0 constitute the well known gauge independent hard thermal loop result which has been previously obtained in the feynman gauge as well as in general covariant gauges @xcite . it also agrees with the results obtained in the static and long - wavelength limits using the temporal gauge with @xmath104 @xcite . from the second lines in eqs . ( [ eq11 ] ) and ( [ eq12 ] ) one can extract the part proportional to @xmath105 which comes from the region of high internal momenta . such contributions can be readily obtained noticing that @xmath106 . in this way , inserting the logarithmic part of eqs . ( [ eq11 ] ) and ( [ eq12 ] ) into eq . ( [ eq7a ] ) , we obtain the following result for the logarithmic contribution to the gluon self - energy in the temporal gauge @xmath107(k_{\mu}k_{\nu}-k^{2}g_{\mu\nu } ) \right . \nonumber \\ + \left.{4\alpha\over 3n^{4}}[(n\cdot k ) k_{\mu}-k^{2}n_{\mu}][(n\cdot k)k_{\nu } -k^{2}n_{\nu}]\right\}\ln t , \end{aligned}\ ] ] where we have substituted @xmath108 and @xmath109 . comparing the above logarithmic contribution with the dimensionally regularized zero - temperature gluon self - energy in the general axial gauge @xcite , we find that the coefficients of @xmath1 and of @xmath110 are identical up to a relative sign . we have shown that , in a class of temporal gauges , the exact thermal self - energy of the gluon is transverse to all orders . ( except for the feynman gauge , where the transversality has been verified only to one loop order , this property is not valid in the class of covariant gauges@xcite . ) the forward scattering amplitude associated with the thermal gluon self - energy has been obtained , to one - loop order , using the approximation which consists in neglecting the contributions from the _ prescription poles_. our explicit calculations of the @xmath0 and @xmath111 terms , together with the arguments in @xcite , show that such corrections affect only the contributions proportional to powers of @xmath9 , which become negligible at high temperature . furthermore , our approach preserves the transversality property of the exact thermal gluon self - energy . the leading @xmath0 and the sub - leading @xmath1 contributions are consistently obtained from the forward scattering amplitude of on - shell thermal particles . the relation between the dimensionally regularized zero temperature gluon self - energy and the @xmath13 contribution , given in eq . ( [ eq13 ] ) , is relevant to explain the cancellation of the @xmath112 terms between the zero temperature and the temperature dependent parts of the gluon self - energy ( such a cancellation has also been verified by explicit calculations in covariant gauges @xcite ) . ftb and jf would like to thank cnpq ( brazil ) for financial support . ftb would like to thank the hospitality of the department of applied mathematics of the university of western ontario ( canada ) where part of this work was completed and d. g. c. mckeon for helpful discussions . frm would like to thank fapesp for financial support . in this appendix we illustrate the forward scattering amplitude technique by considering the following type of integral which arises in the calculation of the two - point function in the imaginary time formalism @xcite @xmath113,\ ] ] where @xmath114 , @xmath115 , and we have factorized the denominators @xmath116 from the propagators . in covariant gauges @xmath117 has no poles in the complex @xmath118-plane , while for temporal gauges eq . ( [ propag ] ) yields extra _ prescription poles_. for the ghost loop diagram in fig 1c we have ( in the temporal gauge ) the trivial situation @xmath119 . the two special cases @xmath120 and @xmath121 corresponds to fig 1a and figs 1b , respectively . in what follows we will consider the case @xmath122 so that there will be contributions from the poles at @xmath123 and @xmath124 . the case @xmath125 can be easily derived using the same basic steps described below . since the integration in ( [ genint ] ) is over all values of @xmath126 , it is convenient to make the change of variables @xmath127 in all the terms @xmath128 so that @xmath129.\ ] ] factorizing the denominators in ( [ genint2 ] ) we can write @xmath130 \end{array}.\ ] ] the @xmath118 integration can be readily performed using the cauchy theorem and closing the contour in the right hand side plane where there are poles at @xmath131 and @xmath132 ( @xmath133 is a pure imaginary quantity at this stage of the calculation ) . in this way we obtain @xmath134 where the subscript @xmath135 is to remind us that we have only considered the _ mass shell poles_. performing the change of variables @xmath136 in the second term of ( [ genint4 ] ) we can write @xmath137 finally , using the property @xmath138 and taking into account the contributions @xmath139 we obtain @xmath140 using @xmath141 , the special case when @xmath142 can be written as @xmath143_{q_0 = |\vec q|}.\ ] ] the expression inside the bracket is a typical contribution to the on - shell forward scattering amplitude as represented , for example , by the first two terms in the eq . ( [ eq6 ] ) . here we show explicitly that the one - loop contributions to @xmath144 and @xmath145 , in the class of temporal gauges characterized by the gauge parameter @xmath14 , are identical to zero . we start from the general expression @xmath146\coth\left(\frac{q_0}{2t}\right),\end{aligned}\ ] ] where the integrands @xmath147 are obtained projecting the contributions from the diagrams in figs . 1a and 1b onto the tensors @xmath148 . using partial fraction decompositions , as in eq . ( [ eq4 ] ) , we perform shifts @xmath149 in all the terms involving @xmath150 in the denominator , which are justified in the dimensional regularization scheme . after using also the property @xmath151 as well as the invariance of the @xmath152-integral under @xmath153 , the rather involved expressions of @xmath154 simplify to @xmath155\ ] ] and @xmath156.\ ] ] it is interesting to note that the terms with denominators involving factors like @xmath157 or @xmath150 cancel after performing shifts , so that we do not need to worry about _ prescription poles_. hence the only poles which may contribute are the ones at @xmath27 and @xmath158 . the corresponding residues of @xmath154 are @xmath159 = \frac{1}{2|\vec k||\vec q| } \left.\left(4 q\cdot u - 4 k\cdot u \frac{k\cdot q}{k^2 } - k\cdot u \right)\right|_{q_0=|\vec q|},\ ] ] @xmath160 = -\frac{1}{2|\vec k||\vec q + \vec k| } \left.\left(4 q\cdot u - 4 k\cdot u \frac{k\cdot q}{k^2 } - k\cdot u \right)\right|_{q_0=|\vec q+\vec k|-k_0}\ ] ] @xmath161 = \frac{1}{2|\vec q| } \left.\left(4 \frac{k\cdot q}{k^2 } - 1 \right)\right|_{q_0=|\vec q|}\ ] ] and @xmath162 = -\frac{1}{2|\vec q+\vec k| } \left . \left(4 \frac{k\cdot q}{k^2 } + 3 \right)\right|_{q_0=|\vec q+\vec k|-k_0}.\ ] ] performing the shift @xmath163 in the eqs . ( [ app5 ] ) and ( [ app7 ] ) one can see that the contributions from the distinct poles in @xmath154 exactly cancel each other .
the approach which relates thermal green functions to forward scattering amplitudes of on - shell thermal particles is applied to the calculation of the gluon self - energy , in a class of temporal gauges . we show to all orders that , unlike the case of covariant gauges , the exact self - energy of the gluon is transverse at finite temperature . the leading @xmath0 and the sub - leading @xmath1 contributions are obtained for temperatures @xmath2 high compared with the external momentum . the logarithmic contributions have the same structure as the ultraviolet pole terms which occur at zero temperature .
bilayer graphene @xcite provides an attractive platform for unconventional two - dimensional electronic physics due to the two quadratic band contacts at its fermi points , and because of the variety of ways by which one can introduce a band gap and produce momentum - space berry curvature @xcite . the interlayer nearest - neighbor hopping , @xmath0 , warps the band structure of the individual graphene layers , repelling two bands away from the fermi energy and leaving the remaining two dispersing quadratically . this warping is a consequence of the two - step process in which electrons hop between the two low - energy sublattices via the two high - energy ones . the band touching points at inequivalent brillouin zone corners @xmath1 or @xmath2 are protected by the berry phase @xmath3 ( or winding number @xmath4 ) and the required chiral ( or sublattice ) symmetry between the low - energy sublattices on opposite layers @xcite . keeping all of these symmetries , the band touching point can only split , instead of being gapped out , even when trigonal warping and other weak remote - hopping processes are taken into account @xcite . however , the chiral symmetry between the low - energy sublattices can be intrinsically broken by spin - orbit coupling @xcite , spontaneously broken by electron - electron interactions @xcite , or explicitly broken by an interlayer potential difference @xcite . as a consequence , the quadratic band touching is no longer symmetry - protected and gaps open up at valleys @xmath1 and @xmath2 . while the first two types of gaps are small in practice @xcite , the electric - field - induced gap , which is the focus of this paper , saturates at a large value comparable to the interlayer hopping @xmath5 ev @xcite . opening the band gap produces large momentum - space berry curvature in the quasiparticle bands , with the curvature integral quantized to @xmath6 over a half brillouin zone centered at @xmath1 or @xmath2 @xcite . moreover , for bilayer graphene gapped by an electric field , the sign of this partial chern number depends on the valley index , the sign of the energy gap ( given by the direction of interlayer electric field ) , and the layer stacking order ( i.e. , ab or ba ) @xcite . here ab ( ba ) stacking refers to the case in which @xmath0 couples the top layer a ( b ) and bottom layer b ( a ) high - energy sublattices . in the presence of an interlayer electric field , when the field direction is reversed across a line @xcite or when the field is uniform and the layer stacking switches from ab to ba @xcite , the valley - projected chern number changes by @xmath7 ( @xmath8 ) across the domain wall in valley @xmath1 ( @xmath2 ) . as a result , both types of domain walls host two chiral edge states in each valley with chirality ( direction ) locked to valley index @xmath1 or @xmath2 , as shown in fig . [ figwall ] . similar domain wall states also occur spontaneously due to interactions in the absence of electric fields but at finite temperature @xcite . importantly @xcite , these `` quantum valley hall '' ( qvh ) edge states are not strictly topological and can be gapped out by a sufficiently strong , large - momentum scattering which couples the two valleys , even if the underlying symmetries are still preserved . it is therefore crucial that valley index also remains a `` good quantum number , '' for which we will assume that short - range disorder , interlayer stacking , and electric field direction changes are smooth on the scale of the lattice . under this assumption , backscattering is prohibited and the system of domain walls provides an attractive platform for tomonaga - luttinger liquid physics @xcite . in this paper , we study the electronic transport properties of a junction where two domain walls intersect ( fig . [ figqpc ] ) . such a structure resembles the quantum point contact of the edges of two quantum spin hall ( qsh ) insulators , which has been studied in refs . and can be probed by four - terminal transport measurements . a domain wall junction can be tuned through a `` pinch - off transition '' by applying a local field ( such as a perpendicular electric field ) to the junction region . in ref . , it was found that the corresponding pinch - off transition for qsh systems is controlled by a novel quantum critical point , and that at low temperatures the conductance is described by a universal scaling function across the pinch - off transition . in contrast to the qsh edge states , which have a single time - reversed pair of helical modes , the domain wall states in bilayer graphene host four helical pairs ( including electron spin ) . we find that this leads to several important modifications of the low - energy properties of the junction . unlike with the qsh edge states , whose forward - scattering interactions are characterized by a single luttinger parameter , it has been argued that for the domain wall states in bilayer graphene , one should characterize interactions with two independent luttinger parameters @xcite . this leads to an expanded phase diagram for the possible stable states of the junction . moreover , we find that the pinch - off transition is modified . depending on the interaction strengths , there are two possible regimes for the reduced , two - terminal conductance : one in which it undergoes a single pinch - off transition directly from @xmath9 to @xmath10 and one in which it undergoes two separate transitions , separated by a stable state with conductance @xmath11 . we study the critical behavior of these transitions and compute the universal crossover scaling functions for weak interactions . this paper is structured as follows . first we introduce in detail the domain wall states in bilayer graphene and derive low - energy effective field theories for them . adding interactions , we show how these states are luttinger liquids described by two independent luttinger parameters . from there , we characterize the geometry of two intersecting domain wall states in the language of the resulting effective quantum point contact . we then analyze the resulting four - terminal junction in both the context of many - body tunneling in a bosonization framework and with an s - matrix renomalization group using diagrammatic perturbation theory . combining these analyses , we determine the behavior of the reduced , two - terminal conductance over a range of interaction strengths . in this paper , we calculate several measurable properties of bilayer graphene domain wall quantum point contacts . in section [ sec : domainwalls ] , we find the critical exponent @xmath12 which characterizes the low temperature tunneling conductance scaling for a single domain wall , a result previously derived in ref . . in [ sec : fourterminal ] , we introduce a diagonal conductance @xmath13 , which is only strictly quantized when valley index is conserved both within individual domain walls and across the junction . this conductance , therefore , stands as a first test of whether experimental samples are in the appropriate disorder regime for the analysis in this paper . we also show in [ sec : fourterminal ] that states of exotic charge and valley tunneling character dominate the conductance of the junction under very strong attractive or repulsive interactions ( fig . [ figphase ] ) . finally , in [ sec : scaling ] , we show that the left - to - right conductance @xmath14 undergoes either a direct transition from @xmath9 to @xmath10 or an indirect one with an intermediate step to @xmath11 depending on experimental specifics ( fig . [ figzerot ] ) . labeling the direct transition @xmath15 and the first step of the indirect transition @xmath16 , we go on to show that at low temperatures , the conductance transitions should collapse onto universal scaling functions @xmath17 with critical exponents @xmath18 as functions of the interaction strengths ( figs . [ figtemp],[figscaling ] ) . in this section we introduce our model system of bilayer graphene domain wall modes . first , we begin with the hamiltonian for a single domain wall and the luttinger liquid physics which govern it in the presence of interactions . then , we discuss the four - terminal geometry which arises at the intersection of two domain walls and its equivalence to a quantum point contact . is near the chiral symmetric point . adopting the notation of ref . , the two domain wall states in each direction are labeled @xmath9 and @xmath19 working from the brillouin zone edge in . when the fermi energy is exactly in the middle of the bulk gap , the fermi velocities are the same for the @xmath9 and @xmath19 bands and electron direction is set by valley index @xmath1/@xmath2.,width=336 ] as discussed in the introduction , bilayer graphene domain walls can be created by varying the direction of the interlayer electric field or by varying the interlayer stacking order . in either case , the valley - projected chern number changes by @xmath7 ( @xmath8 ) across the domain wall in valley @xmath1 ( @xmath2 ) . this necessitates the existence of two domain wall states in each valley , with the states at @xmath1 having equal and opposite velocities to those at @xmath1. adopting the notation of ref . , we label these bands @xmath9 and @xmath19 respectively ( fig . [ figwall ] ) . for the purposes of our model , we will assume that the fermi energy lies exactly in the middle of the bulk gap , which allows the simplification @xmath20 . this allows us to write down the non - interacting hamiltonian density @xmath21 where the indexes `` in '' and `` out '' refer to direction and can correspond to electron operators with valley index @xmath1 or @xmath2 depending on the orientation of the domain wall . we will see later that in a four - terminal structure , the correspondence between in / out and @xmath22 will alternate with lead index due to the valley - momentum locking of the domain wall modes . in the limit of the fermi energy resting exactly at the chiral - symmetric point , the band indexes become arbitrary labels for all of the calculations in this paper . a consequence of this is the emergence of a band - index - exchange symmetry in this problem , which we will frequently highlight in calculations throughout this paper . deviations from this point in the fermi energy are expected in physical systems , and will lead in general to a relaxation of this symmetry . if the deviations are small , the physics should resemble the predictions of this paper , with corrections of order @xmath23 . these corrections greatly complicate the calculations in this paper and obscure key generalities , and to that end we consider @xmath20 a desirable simplification of this problem . calculations by killi , wei , affleck , and paramekanti indicated that for interacting bilayer graphene systems , the interaction is dominated by two density - density interactions @xcite : @xmath24 @xmath25 is the usual two - body forward scattering term which leads to luttinger liquid physics and @xmath26 is a new one which breaks the u(2 ) symmetry of electron distribution between the @xmath9 and @xmath19 bands . both can be effectively tuned by altering the strength of the perpendicular electric field , though for all reasonable numerical estimates , ref . found @xmath27 and @xmath28 harder to tune , which is sensible as only @xmath25 contains contributions from the long - range part of the coulomb interaction . other density - density interactions , specifically those which affect electron spin , should be small in practice . in the absence of an external magnetic field , and given the weak spin - orbit interaction in graphene , electron spin should remain su(2)-symmetric . in this limit , the electron spin sectors of this problem should remain noninteracting and terms which lead to phenomena such as spin - density waves will be marginally irrelevant @xcite . returning to our hamiltonian , we can consider a single domain wall by bosonizing , @xmath29 where @xmath30 ; @xmath31 ; @xmath32 ; and @xmath33 is the short wavelength cutoff . the bosonic fields @xmath34 obey the commutation algebra : @xmath35=i\pi\delta_{ab}\delta_{\sigma\sigma'}\tau^{z}_{ij}\operatorname{sgn}(x - y ) . \label{ogcommutator}\ ] ] under this transformation , the bare hamiltonian and interactions become : @xmath36 \nonumber \\ v_{\pm } & = & \frac{\hbar v_{f}}{8\pi}\lambda_{\pm}\sum_{\sigma=\uparrow,\downarrow}\big[(\partial_{x}\phi_{0\sigma , in } - \partial_{x}\phi_{0\sigma , out } ) \nonumber \\ & \pm & ( \partial_{x}\phi_{\pi\sigma , in } - \partial_{x}\phi_{\pi\sigma , out})\big]^{2 } \label{hint}\end{aligned}\ ] ] where @xmath37 . the interacting hamiltonian can be simplified by the sum / difference changes of basis : @xmath38 where the @xmath39 and @xmath40 sectors are charge and spin respectively . in this basis , all of the interactions are in the charge sector and , as motivated earlier in this section , the spin sector is noninteracting : @xmath41 \nonumber \\ v_{\pm } & = & \frac{\hbar v_{f}}{8\pi}\lambda_{\pm c}\left[\partial_{x}\phi_{\pm c , in } - \partial_{x}\phi_{\pm , out}\right]^{2}.\end{aligned}\ ] ] the plus and minus charge sectors of @xmath42 are then each renormalized by only @xmath43 respectively , encouraging us to express the interaction parameter @xmath44 separately for each charge sector . therefore we can write down the interacting hamiltonian for each charge sector , @xmath45 . diagonalizing this hamiltonian , the definition of the luttinger parameters @xmath46 arises naturally . the change of basis @xmath47 returns our interacting hamiltonian to the form of one for non - interacting chiral bosons @xmath48\end{aligned}\ ] ] where @xmath49 in the new basis , the charge fields @xmath50 obey the commutation relation : @xmath51=i\pi g_{u}\delta_{uv}\tau^{z}_{ij}\operatorname{sgn}(x - y)\ ] ] where @xmath52 ; @xmath32 ; and we note that the noninteracting spin sector fields still obey this commutation relation with @xmath53 . as in refs . , the tunneling density of states for a single edge @xmath54 is controlled by the interactions . however here , unlike in the qsh case , the critical exponent is a function of two luttinger parameters , such that in agreement with ref . , @xmath55 from an experimental perspective , measuring this critical exponent for the tunneling conductance would be a valuable first step in confirming the luttinger liquid physics of these bilayer graphene domain wall states . are lead indexes and the two modes displayed for each domain wall are those at @xmath9 and at @xmath19 . all of the modes shown here are at valley @xmath1 ; a counterpropagating set of modes exists at @xmath2 and is related by time - reversal symmetry . including electron spin , there are 4 modes in each valley in each domain wall , for a total of 16 modes to consider for this qpc structure.,width=336 ] a pattern of two domain walls which pass nearly by each other can be formed by varying either electric field direction or interlayer stacking twice ( fig . [ figqpc]a ) . if we distort the central region of this picture , we could imagine bringing the two domain walls so close that tunneling between them becomes significant . in this case , the two domain walls have formed the equivalent of a quantum point contact ( fig . [ figqpc]b ) . mapping our work in [ sec : domainwalls ] onto a spinful luttinger liquid for the two - wall system , we deduce in this section , for all interaction regimes , the interaction strengths at which many - body tunneling processes become relevant and alter the junction s charge and valley conductances . while each lead retains the properties and luttinger parameters from [ sec : model ] individually , we will find it convenient to limit the usage of the @xmath56 basis to the treatment of isolated domain walls and adopt a new basis with effective charge and valley sectors . the charge sector which arises here is a new degree of freedom and comes from a rotation of the indexes for @xmath1 and @xmath2 and propagation direction . we will label _ these _ charge and valley sectors @xmath57 and @xmath58 respectively to differentiate them from the charge and spin sectors which arose in [ sec : domainwalls ] , which are labeled @xmath39 and @xmath40 respectively . the two domain walls in fig . [ figqpc ] have opposite helicties , due to being on the top ( bottom ) of the central region . we can define , for the interacting system , fields labeled by sum / difference , charge / spin , direction , and valley ( @xmath59 ) : @xmath60 where @xmath61 is the heaviside step function and the indexes @xmath62 on the noninteracting @xmath63 refer to the individual lead in figure [ figqpc ] with which they are associated . the intersection of the two domain walls occurs at @xmath64 such that our theory consists of four , isolated domain walls everywhere except at that point . assuming that the interaction strength is controlled globally such that each lead has the same values of @xmath65 , we can create a @xmath66 basis for each @xmath67 sector of the combined hamiltonian of the two domain walls @xmath68 : @xmath69 \nonumber \\ \phi_{\pm c / s , lk } & = & \frac{1}{2}\left[\phi_{\pm c / s,\rho } + \phi_{\pm c / s , v } - \theta_{\pm c / s,\rho } - \theta_{\pm c / s , v}\right ] \nonumber \\ \phi_{\pm c / s , rk ' } & = & \frac{1}{2}\left[\phi_{\pm c / s,\rho } - \phi_{\pm c / s , v } + \theta_{\pm c / s,\rho } - \theta_{\pm c / s , v}\right ] \nonumber \\ \phi_{\pm c / s , lk ' } & = & \frac{1}{2}\left[\phi_{\pm c / s,\rho } - \phi_{\pm c / s , v } - \theta_{\pm c / s,\rho } + \theta_{\pm c / s , v}\right ] \nonumber \\ \end{aligned}\ ] ] where again @xmath70 are the charge and spin sectors which resulted from rotating the indexes for @xmath71 and @xmath72 are the indexes which have , along with the choice of @xmath73 , resulted from rotating indexes for propagation direction ( @xmath74 ) and valley ( @xmath59 ) . the new fields are governed by the modified commutation relation : @xmath75=2\pi i\delta_{uv}\delta_{\alpha\beta}\delta_{ij}\theta(x - y)\ ] ] where @xmath52 ; @xmath76 ; and @xmath77 . this unitary rotation of the variables effectively changes the sign of the interaction `` cross - term '' individually for the choice of @xmath78 , @xmath79 , @xmath57 , and @xmath58 within the @xmath39 sector : @xmath80 \nonumber \\ & -&\lambda_{\pm}\bigg[(\partial_{x}\phi_{\pm c\rho})(\partial_{x}\phi_{\pm c\rho } ) - ( \partial_{x}\phi_{\pm cv})(\partial_{x}\phi_{\pm cv } ) \nonumber \\ & - & ( \partial_{x}\theta_{\pm c\rho})(\partial_{x}\theta_{\pm c\rho } ) + ( \partial_{x}\theta_{\pm cv})(\partial_{x}\theta_{\pm cv})\bigg]\bigg\}.\nonumber \\ \label{spinful}\end{aligned}\ ] ] the previous equation , though diagonal , was left unsimplified and in the form of eq . [ hint ] such that by the same logic as in [ sec : domainwalls ] , the form of the simplified diagonalized hamiltonian , as well as the interactions , can just be read off : @xmath81 such that @xmath82 now has the form of a spinful luttinger liquid . in this basis , both the interacting and noninteracting hamiltonians are diagonal and so the transformation between the interacting and noninteracting @xmath83 requires just a simple rescaling by the interaction parameter for each sector . for this geometry , one can probe experimentally by measuring the current @xmath84 at one of the leads in response to an applied voltage on another lead @xmath85 such that a @xmath86 conductance matrix characterizes the system , @xmath87 where @xmath88 is a lead index . in the presence of time - reversal and valley symmetries , the number of independent or nonzero parameters in @xmath89 is greatly reduced , as described in detail in the appendix of ref . . for this system , we can then consider a reduced set of voltages and currents : @xmath90 where @xmath91 is the left - to - right current and @xmath92 is the top - to - bottom current . @xmath93 and @xmath94 are similarly defined such that @xmath93 is a bias of leads @xmath95 and @xmath96 relative to leads @xmath7 and @xmath97 and @xmath94 is a bias of leads @xmath95 and @xmath7 relative to leads @xmath97 and @xmath96 . therefore @xmath14 and @xmath98 are the two - terminal conductances measured left - to - right and top - to - bottom respectively . @xmath99 are skew conductances , equal as a consequence of time - reversal symmetry . in the noninteracting model , this skew conductance is zero as a consequence of artificial spatial symmetries , such as mirror symmetry . though it may become nonzero under increased interaction strengths , the skew conductance is still negligible along the relevant directions which characterize transitions in this system @xcite . we can define a final current across the junction @xmath100 , which one can probe by applying a voltage @xmath101 which biases leads @xmath95 and @xmath97 relative to leads @xmath7 and @xmath96 , with a conductance @xmath102 if valley is conserved , then electrons can not enter at lead @xmath95 and exit at lead @xmath97 , implying that a measurement of an exactly quantized @xmath103 would be an experimental confirmation that valley - nonconserving disorder is absent and the system is appropriately described by the physics in this paper . the factor of @xmath104 in the landauer prediction @xmath105 comes from factors of 2 for band index ( @xmath9 and @xmath19 ) , electron spin degeneracy , and the two incoming leads at @xmath1 ( @xmath95 and @xmath97 ) . we can also , in a similar manner , characterize the valley conductance of the system in terms of left - to - right and top - to - bottom parameters @xmath106 and @xmath107 . before the introduction of any interactions or tunneling operators , our system consists of two , left - to - right domain wall states and we consider it `` fully - open . '' for this system , @xmath108 and @xmath109 such that it is a left - to - right _ charge conductor , valley conductor _ , which we will denote as the cc phase . a @xmath110 rotation and relabeling ( with regards to fig . [ figqpc ] ) or the pinch - off inversion of this phase , for which @xmath111 and @xmath112 , is considered `` fully pinched - off '' and is a left - to - right _ charge insulator , valley insulator _ , which we denote as the ii phase . with this framework established , we can examine perturbatively tunneling processes between the two adjacent domain walls which may lead to differing charge and valley conductances . using our bosonization work , we can examine the rescaling of the coupling strength for each process , noting the interaction regime in which it dominates the physics of the quantum point contact . figure [ figtv ] illustrates the single - particle valley - conserving processes which can exist in this system within a single spin channel . in general , many - body tunneling processes will also be present and may become relevant . these many - body tunneling processes can be considered products of single - electron - tunneling processes which , in the most general case , may or may not conserve spin or valley indexes . however , restricting ourselves to the set of processes which conserve valley , it becomes apparent that the linear combinations of bosonized operators which can lead to relevant operators can only be achieved through products of single - particle processes which conserve spin . therefore , in the analysis of many - body processes which may become relevant and drive to phases with different conductance behavior , we can simply consider products of spin- and valley - conserving single - electron tunneling : @xmath113 where @xmath114 ; @xmath31 ; @xmath115 ; @xmath116 is the coupling strength of the process ; and @xmath117 is an arbitrary valley- and spin - conserving single - particle tunneling process . for the sake of condensing notation , tunneling from @xmath118 will be expressed as @xmath119 . weak tunneling about the cc phase ( fig . [ figtv]a ) is related to weak backscattering in the left - to - right direction , and is also dual to weak tunneling about the ii phase ( fig . [ figtv]b ) . taking advantage of this duality , we will restrict our discussion to the set of @xmath58 operators which may destabilize the cc phase , noting that the @xmath120 tunneling operators about the ii phase are related by a duality . this duality is explored in greater detail in ref . . ( [ scalingd1 ] ) and ( [ scalingd3 ] ) ) and the fully open ( cc ) or pinched - off ( ii ) junction phases are destabilized . the central dot at @xmath121 is the noninteracting point and the dotted oval is the region of predicted accessible interaction strength in ref . for a suspended sample . the ic ( ci ) regions are characterized by four - body tunneling processes which transmit exclusively valley ( charge ) across the junction . the b regions represent relevant eight - body tunneling processes which are charge insulating ( @xmath122 ) or conducting ( @xmath123 ) and differ by band - index and valley - transmission character . regions of overlap between boundaries , denoted with @xmath124 , have multiple relevant operators at different orders and presumably more complicated behavior . in the central region , the fully open ( cc ) or pinched - off ( ii ) phases remain stable and the conductance is characterized by single - electron tunneling.,width=336 ] the near - intersection of the two domain walls is a @xmath125-dimensional object , and therefore the coupling strength @xmath126 of a given tunneling process flows , to first order , as @xmath127 where @xmath128 is the scaling dimension of the tunneling process @xmath129 . to understand which operators may become relevant , we can first examine the single - electron tunneling processes ( fig . [ figtv]a ) : @xmath130 where @xmath114 ; @xmath31 ; @xmath115 ; and we have restricted ourselves to processes which preserve spin and valley . for single - particle tunneling , @xmath131\ ] ] such that single - electron tunneling is always marginal or irrelevant ( @xmath132 ) for all possible inter- and intra - band scattering processes . in the nearly noninteracting regime , where the _ least irrelevant _ operators are @xmath133 , the strength of @xmath134 can be controlled by an external parameter , such as the gate voltage @xmath135 for a given set of interaction strengths @xmath46 . in this interaction regime , we know that the cc and ii phases are stable and that at least one quantum critical point exists to mediate the transition between them . in the subsequent section , section [ sec : pinchoff ] , we will use diagrammatic perturbation theory in the interactions about the cc phase to search for the set of possible intermediate phases and quantum critical points which characterize the single - electron - tunneling behavior of the junction . the factor of @xmath136 in @xmath137 is due to @xmath138 and will remain an obstacle to a process becoming relevant unless the operator in bosonized form only contains @xmath39-sector variables . for higher - body tunneling , this factor is greater than or equal to @xmath95 , and prevents any process which is nt pairwise spin conserving and invariant under arbitrary su(2 ) spin rotations from becoming relevant . one can view this as pairwise spin conservation allowing decomposition into products of @xmath139 and su(2 ) invariance providing the necessary linear combinations of @xmath31 to isolate @xmath39 sector variables . restricting to processes which conserve valley and obey these spin index constraints , the first operators to become relevant as interactions are increased are therefore a specific set of four- and eight - body tunneling processes . figure [ figphase ] details the region in interaction space for which each class of many - body operators becomes relevant ( @xmath140 ) . when a process becomes relevant , the bosonized operators will become locked into the values which minimize the tunneling operator and become gapped out , altering the conductance of the junction . as an applied voltage only couples to the total electron density , only relevant operators containing @xmath141 or @xmath142 can cause the junction to become charge insulating . in the central region , all operators are marginal or irrelevant , though single - electron tunneling @xmath133 is only the least irrelevant operator close to the non - interacting point @xmath121 . four groups of four - body processes are present which can become relevant and drive to phases which are completely charge or valley insulating : @xmath143 \nonumber \\ & + & v_{ic_{a}}^{(2)}\left[\mathcal{o}^{0\pi}_{\uparrow k}\mathcal{o}^{\pi 0}_{\uparrow k'}\mathcal{o}^{0\pi}_{\downarrow k}\mathcal{o}^{\pi 0}_{\downarrow k'}+ 0\leftrightarrow\pi\right]+\ h.c . \nonumber \\ \nonumber \\ v_{ic_{b } } & = & v_{ic_{b}}^{(1)}\left[\mathcal{o}^{00}_{\uparrow k}\mathcal{o}^{\pi\pi}_{\uparrow k'}\mathcal{o}^{00}_{\downarrow k}\mathcal{o}^{\pi\pi}_{\downarrow k'}+ 0\leftrightarrow\pi\right ] \nonumber \\ & + & v_{ic_{b}}^{(2)}\left[\mathcal{o}^{0\pi}_{\uparrow k}\mathcal{o}^{0\pi}_{\uparrow k'}\mathcal{o}^{0\pi}_{\downarrow k}\mathcal{o}^{0\pi}_{\downarrow k'}+ 0\leftrightarrow\pi\right ] + \ h.c . \nonumber \\ \nonumber \\ v_{ci_{a } } & = & v_{ci_{a}}^{(1)}\left[\mathcal{o}^{00}_{\uparrow k}\mathcal{o}^{00\dagger}_{\uparrow k'}\mathcal{o}^{00}_{\downarrow k}\mathcal{o}^{00\dagger}_{\downarrow k'}+ 0\leftrightarrow\pi\right ] \nonumber \\ & + & v_{ci_{a}}^{(2)}\left[\mathcal{o}^{0\pi}_{\uparrow k}\mathcal{o}^{\pi 0\dagger}_{\uparrow k'}\mathcal{o}^{0\pi}_{\downarrow k}\mathcal{o}^{\pi 0\dagger}_{\downarrow k'}+ 0\leftrightarrow\pi\right ] + \ h.c . \\ \nonumber \\ v_{ci_{b } } & = & v_{ci_{b}}^{(1)}\left[\mathcal{o}^{00}_{\uparrow k}\mathcal{o}^{\pi\pi\dagger}_{\uparrow k'}\mathcal{o}^{00}_{\downarrow k}\mathcal{o}^{\pi\pi\dagger}_{\downarrow k'}+ 0\leftrightarrow\pi\right ] \nonumber \\ & + & v_{ci_{b}}^{(2)}\left[\mathcal{o}^{0\pi}_{\uparrow k}\mathcal{o}^{0\pi\dagger}_{\uparrow k'}\mathcal{o}^{0\pi}_{\downarrow k}\mathcal{o}^{0\pi\dagger}_{\downarrow k'}+ 0\leftrightarrow\pi\right ] + \ h.c . \nonumber \\\end{aligned}\ ] ] or in bosonized form : @xmath144 \nonumber \\ v_{ic_{b } } & = & \cos(\theta_{+c\rho})\left[v^{(1)}_{ic_{b}}\cos(\theta_{-cv } ) + v^{(2)}_{ic_{b}}\cos(\phi_{-c\rho})\right ] \nonumber \\ v_{ci_{a } } & = & \cos(\theta_{+cv})\left[v^{(1)}_{ci_{a}}\cos(\theta_{-cv } ) + v^{(2)}_{ci_{a}}\cos(\phi_{-c\rho})\right ] \nonumber \\ v_{ci_{b } } & = & \cos(\theta_{+cv})\left[v^{(1)}_{ci_{b}}\cos(\theta_{-c\rho } ) + v^{(2)}_{ci_{b}}\cos(\phi_{-cv})\right ] \nonumber \\\end{aligned}\ ] ] where @xmath145 are the coupling constant strengths for the two different choices of interband scattering for each class of tunneling process and have absorbed factors of @xmath7 during bosonization and simplification . the scaling dimensions for these couplings strengths are : @xmath146 in the regions where the ic ( ci ) operators become relevant , the system will be charge ( valley ) insulating and valley ( charge ) conducting . expressed in terms of conductance elements , the ic ( ci ) phases will have @xmath147 , @xmath148 ( @xmath149 , @xmath150 ) . these phases are related to the charge and spin insulating phases for the topological insulator qpc @xcite . however , unlike those phases , the ic and ci phases in the bilayer graphene junction will still be transmitting in the spin sector as long as @xmath151 are relatively close to noninteracting . the remainder of the ii / cc region is bounded by four eight - body processes . each one can be considered as a product of one of the terms in the previous four - body processes multiplied by a product of selected conjugates of itself as to isolate just a single charge - sector variable . in bosonized variables , these eight - body tunneling operators are : @xmath152 the strengths of these operators have scaling dimensions : @xmath153 they are only relevant under extremely strong interactions ( @xmath154 or @xmath155 ) and are charge insulating ( @xmath122 ) or conducting ( @xmath123 ) . higher - order many - body tunneling processes are of course also possibly relevant , however due to the nature of the scaling dimension calculation , they will become relevant at much larger values of interaction strength than the boundaries of the shaded regions in figure [ figphase ] . as demonstrated in the previous section , under weak interactions the junction is stable in either the open ( cc ) phase or the closed - off ( ii ) phase , both of which are characterized by single - electron tunneling and are related to each other by both @xmath110 rotations and the pinch - off duality . in this section , we expand perturbatively in the interactions about the cc fixed point in search of the quantum critical point(s ) which characterize the cc@xmath156ii quantum phase transition . in the process , we discover that , in addition to the @xmath157 critical point , which is expected as a consequence of the pinch - off duality , an additional family of intermediate critical points and phases are also present . for each of the possible paths between the ii and cc phases we derive the conductance signatures which characterize the low - temperature transitions as functions of the two interaction strengths and the external gate voltage . first , we show how the general s - matrix characterizing the junction is renormalized by weak interactions , deriving a phase diagram and renormalization group ( rg ) in the case where scattering between the @xmath9 and @xmath19 bands is disallowed . we then allow for interband scattering , as might be present in the case of relatively smooth disorder , and introduce an s - matrix parameterization incorporating the additional system parameters . using the results of an extensive renormalization group calculation , detailed in appendix [ appendix : stability ] , we assert that the most general s - matrix flows back to one with small - momentum conservation . the rg flow on this surface therefore contains all of the characteristic non - trivial quantum critical points for the pinch - off transition of this problem . finally , we derive the critical exponents and universal scaling functions for the two classes of conductance transitions , up to leading order in the interactions . in the absence of interactions , tunneling through the junction can be characterized by an s - matrix restricted only by time - reversal symmetry and valley - index conservation @xmath158 where @xmath159 are lead indexes @xmath62 ; @xmath160 are spin indexes @xmath71 ; and @xmath161 are band indexes @xmath9 and @xmath19 . given the negligible spin - orbit coupling in graphene , we can consider here that the time - reversal operator @xmath162 leaves the spin sector unaffected , such that @xmath163 . therefore , time - reversal symmetry restricts that @xmath164 . additionally , to keep the domain wall states gapless , one must disallow scattering from @xmath1 to @xmath2 , which restricts elements of the s - matrix @xmath165 . in this section , we will work the most general allowed s - matrix down to one which is characterized by parameters which have physical meaning . beginning with the modes in a single valley @xmath1 : @xmath166 where the rows and columns of @xmath167 indicate scattering of the incoming modes with valley index @xmath1 ( leads @xmath95 and @xmath97 ) to the outgoing ones ( leads @xmath7 and @xmath96 ) . the matrices @xmath168 and @xmath169 live in the @xmath170 space of band indexes . identical copies of @xmath167 exist for the up and down spins . the elements of @xmath167 are otherwise unconstrained if we allow scattering between the @xmath9 and @xmath19 bands such that @xmath167 is an arbitrary @xmath171 matrix . we can choose to parameterize @xmath172 @xmath173 where the @xmath174 ( @xmath175 ) are for now unconstrained u(2 ) matrices which characterize operations on the outgoing ( incoming ) electronic wavefunction at lead @xmath176 ( @xmath177 ) and valley index @xmath1 . in this parameterization , we can choose the tunneling probabilities for each band to be real such that @xmath178 . at this point , @xmath167 remains characterized by 16 free parameters . choosing to parameterize the @xmath174 in terms of euler angles : @xmath179 corresponds to the processes which break small - momentum conservation and scatter electrons between the @xmath9 and @xmath19 bands at lead @xmath176 . @xmath180 corresponds to the scattering phase acquired tunneling from lead @xmath177 to lead @xmath176 at valley @xmath1 . while there are four independent @xmath181 angles , there are only three independent @xmath182 as only the relative scattering phase matters , such that @xmath183 . we have only displayed modes at @xmath1 ; an additional copy of this picture exists for modes at @xmath2 , related by time - reversal symmetry . this picture is valid for electrons with either spin , as our model system is su(2 ) spin - invariant for all values of charge - sector interaction strengths.,width=336 ] recognizing that @xmath184 and @xmath185 are just u(1)@xmath186u(1 ) transformations at each lead , we can gauge them out . we are then left with nine variables which have physical significance . the four @xmath187 mixing angles @xmath188 correspond to the breaking of small - momentum - conservation at lead @xmath176 , the three independent linear combinations of scattering phases @xmath180 each correspond to the phase acquired for electrons scattering from lead @xmath177 to lead @xmath176 at valley index @xmath1 , and the two real tunneling probabilities @xmath189 and @xmath190 characterize the extent to which the junction is pinched off for each band . these seven linearly independent angular variables ( illustrated in fig . [ figvars ] ) and two tunneling probabilities completely span the space of the gauge - independent , time - reversal - symmetric , spin- and valley - index - conserving problem with s - matrix , @xmath191 which has rows and columns characterized by lead indexes @xmath159 . this parameterization of the s - matrix in lead and band - index spaces is restated more explicitly in the beginning of appendix [ appendix : thetaphi ] . in the subsequent sections , we will show how the surface where @xmath192 is not just a welcome simplification of the problem , but also the surface which contains all of the characteristic quantum critical points and their single relevant eigenvectors . as discussed in [ sec : domainwalls ] , this system is characterized by two kinds of density - density interactions . here , we relate the s - matrix to the single - particle green s function and then use diagrammatic perturbation theory in those two interactions to find the leading order corrections to the s - matrix and derive rg flow equations for our system parameters @xcite . we calculate the relevant physics on the high - symmetry surface where the interband scattering angles @xmath193 . in appendix [ appendix : stability ] , we further calculate the rg flow for all nine s - matrix parameters , demonstrating that the angles @xmath194 either flow back to this surface or are marginal and trivial at the quantum critical points . scattering processes from one lead , band , and spin to another can be considered in terms of a single - electron thermal green s function @xmath195\right\rangle\ ] ] where @xmath196 denotes imaginary time ordering and indexes @xmath197 . in the absence of interactions @xmath198 where @xmath199 and the rows and columns of @xmath200 are the @xmath201 indexes such that elements proportional to @xmath202 correspond to @xmath203 . restricting ourselves to the interactions introduced in [ sec : domainwalls ] ( eq . ( [ bosonints ] ) ) , we can use diagrammatic perturbation theory to calculate the leading order corrections to @xmath204 in the presence of weak @xmath205 , such that the luttinger parameters @xmath206 and terms are kept up to @xmath207 . the one - loop corrections to the s - matrix are qualitatively similar to those in ref . , with special care taken to properly sum at each vertex over the tensor structure of the interaction @xmath208 where @xmath209 are band indexes . illustrated in fig . [ figdiagram ] , only two diagrams contribute to the renormalization of the single - particle green s function : @xmath210 with @xmath211 - \sum_{kl}\left(s_{ik}\sigma_{a}s^{\dagger}_{lk}\sigma_{b}s_{lj}\right)^{\alpha\beta}\operatorname{tr}[\sigma_{a}s^{\dagger}_{lk}\sigma_{b}s_{lk } ] \right]\ ] ] where @xmath212 and @xmath213 are the ultraviolet and infrared cutoffs respectively . traces refer to band index space , @xmath214 , and @xmath215 ( @xmath216 ) ; tracing over the spin degree of freedom is implicit , which leads to the prefactor of @xmath7 . the two terms correspond to the diagrams ( a ) and ( b ) respectively in fig . [ figdiagram ] . we can derive flow equations for the elements of @xmath202 by rescaling the cutoff @xmath217 , perturbation theory . note that @xmath218 are spatial lead indexes and @xmath219 are band indexes which are summed at each vertex over @xmath220 . @xmath221 are spin indexes for which the @xmath222 within each green s function has already been taken into account . even though the system is spin invariant , the spin index on the loop , here @xmath223 , must still be summed over its two values to calculate the physical rg flow . each diagram contributes a logarithmic correction to the s - matrix which , when renormalized , leads to a term in eq . ( [ sflow]).,width=336 ] @xmath224 - \sum_{kl}\left(s_{ik}\sigma_{a}s^{\dagger}_{lk}\sigma_{b}s_{lj}\right)^{\alpha\beta}\operatorname{tr}[\sigma_{a}s^{\dagger}_{lk}\sigma_{b}s_{lk } ] \right ] . \label{sflow}\ ] ] one can immediately observe that this implies that @xmath193 is a fixed surface , since perturbative corrections to @xmath225 require multiplying by @xmath226 in the above equations . we can thus , for now , simplify our focus to the fixed surface where @xmath227 . in appendix [ appendix : stability ] , we calculate the stability of the quantum critical points on this surface for a general set of @xmath181 and @xmath182 , and demonstrate that for physical values of @xmath228 relative to @xmath229 , the critical points are stable in all possible out - of - plane , @xmath230- and valley - symmetric directions . using our parameterization of @xmath202 on the @xmath227 surface , we can exploit the matrix structure of the diagrammatic perturbations to obtain flow equations for the transmission probabilities for each band @xmath231 \nonumber \\ \frac{dt_{\pi}}{dl } & = & -2t_{\pi}(1-t_{\pi})\bigg[(\epsilon_{+ } + \epsilon_{-})^{2}(1 - 2t_{\pi } ) \nonumber \\ & + & ( \epsilon_{+ } - \epsilon_{-})^{2}(1 - 2t_{0})\bigg ] . \label{floweq}\end{aligned}\ ] ] this system of equations obeys two key symmetries . first and foremost , like the qsh point contact described in ref . , it is invariant under the pinch - off duality @xmath232 , which we introduced in [ sec : fourterminal ] . ( [ floweq ] ) are also invariant under exchange of the band indexes @xmath9 and @xmath19 . in the context of our renormalization group calculation , @xmath9 and @xmath19 are arbitrary labels for the band degree of freedom , and so even though this system contains stable fixed points with broken band - index symmetry , the system of flow equations itself must be band - index symmetric . graphically , these two symmetries manifest themselves as two mirror symmetries in fig . [ figquad ] : one about @xmath233 and the other about @xmath234 . , which control the pinching off of the junction , calculated to quadratic order in the interactions @xmath235 ( eq . ( [ floweq ] ) ) , panels ( a)-(c ) . large circles correspond to stable fixed points and small circles indicate nontrivial quantum critical points . the flow is controlled by the ratio of the interaction strengths @xmath236 . when @xmath237 , the @xmath9 and @xmath19 bands are completely decoupled and each one behaves individually as a copy of the qsh problem in ref . ( a ) . for @xmath238 , a set of intermediate fixed points exists which allows the @xmath9 and @xmath19 bands to be pinched off independently ( b ) . when @xmath239 , the quadratic theory predicts that a fixed line will exist @xmath233 ( c ) . higher - order corrections about this line , calculated in appendix [ appendix : bosonization ] , infer flow along it back to the central quantum critical point @xmath240 ( d).,width=336 ] this system of flow equations can have as many as nine fixed points to quadratic order in the interactions ( fig . [ figquad ] ) . the two corner fixed points at @xmath241 are stable for all values of @xmath242 . the central point at @xmath240 controls the transition between the cc and ii corners and is related to the @xmath243 critical point in the single - band ti qpc case @xcite . as in that case , the existence of this central quantum critical point is mandated by the pinch - off duality @xmath244 . the corner fixed points at @xmath245 and @xmath246 represent new , stable , single - electron - tunneling phases where _ only one of the bands _ is pinched off . we label these intermediate , mixed - band - character fixed points as `` m phases '' . transitions between the fully open or closed phases and these m phases are controlled by four fixed points which exist for @xmath238 ( figs . [ figquad]a , [ figquad]b ) . for @xmath247 , the @xmath9 and @xmath19 bands are completely decoupled and each one acts as a ( spin - doubled ) independent copy of the qsh problem in ref . [ figquad]a ) . when @xmath248 , eq . ( [ floweq ] ) predicts that all of the intermediate transitions and stable fixed points will collapse onto a fixed line at @xmath233 ( fig . [ figquad]c ) . this @xmath239 case represents a restoration of u(2 ) band - index symmetry locally at each lead under which the band indexes become trivial . we can examine this fixed line in greater detail by expanding upon our bosonization calculations from [ sec : fourterminal ] . in appendix [ appendix : bosonization ] , we increase the strength of @xmath249 single - electron tunneling to drive from the cc phase towards an action about the m - phase corner fixed points for @xmath250 . calculating higher - order correlation functions about this theory and expanding perturbatively again in the interactions , we discover additional fixed points : @xmath251 where the second point is implied by the combination of mirror reflections about the pinch - off and band - index - exchange lines . taking @xmath252 , this theory exhibits flow back towards the central quantum critical point @xmath240 ( fig . [ figquad]d ) . the simplest assumption would be to postulate that , to lowest order , this flow continues away from the vicinity of the m points without additional fixed points appearing . this implies that the fixed line at @xmath239 is simply an artifact of the @xmath207 perturbation theory and that for extremely small @xmath236 , the m phase is unstable and flow lines in that region point towards the central quantum critical point @xmath240 . from this information , we obtain figure [ figrg ] , a schematic which incorporates the quadratic - order rg flow and the higher - order corrections near the @xmath233 line . as highlighted by the red and purple arrows in that figure , each of the two classes of nontrivial quantum critical points is characterized by only a single relevant direction in the @xmath253 plane . all other out - of - plane @xmath181 and @xmath182 directions are irrelevant or trivially marginal ( appendix [ appendix : stability ] ) . for each class of quantum critical point , we can , knowing that its only relevant eigenvector lies on the @xmath227 plane , use the quadratic - order flow equations to calculate the universal conductance scaling and critical exponents to leading order in the interactions . returning to a discussion of reduced conductance matrixes from [ sec : fourterminal ] , we can express the left - to - right , two - terminal conductance @xmath14 in terms of the s - matrix . the elements of the four - terminal conductance @xmath254 in the lead basis are related to the s - matrix by : @xmath255\ ] ] where @xmath159 are lead indexes @xmath62 such that the matrix @xmath256 plane , combining information from eq . ( [ floweq ] ) and appendix [ appendix : bosonization ] . there are two classes of quantum critical points . the central point controls transitions between the fully open ( cc ) phase and the fully pinched - off ( ii ) phase . four additional critical points on the edges control transitions between the cc / ii phases and an intermediate mixed ( m ) phase in which the two bands have differing conductance contributions . the width of the m phase is @xmath257.,width=336 ] where @xmath258 and the factor of @xmath7 on the conductance is due to electron spin degeneracy . the linear combinations of lead currents @xmath259 which give @xmath260 are a result of , in combination with the requirement @xmath261 , @xmath262 where @xmath263 such that @xmath264 this confirms the result from [ sec : fourterminal ] that , for the valley - conserving problem , @xmath265 is quantized to be @xmath10 regardless of the junction state . this reduction also confirms that , in terms of the s - matrix elements , @xmath266 with the structure of the conductance matrix established we can analyze the finite - temperature conductance transitions near each transition voltage @xmath267 . first , we will consider the direct ii - cc phase transition , for which @xmath14 scales from @xmath9 to @xmath10 . we can write the conductance in its scaling form @xmath268 where @xmath269 in fig . [ figzerot ] , @xmath39 is a non - universal constant , and the subscript @xmath15 on @xmath270 and @xmath271 denotes the direct quantum phase transition between the ii and cc phases . observing that infinitesimal movement of @xmath272 away from @xmath9 is irrelevant ( fig . [ figrg ] ) , we can set @xmath273 and characterize the conductance transition from ii - cc with a single parameter , @xmath274 this equation can be integrated to determine the crossover scaling function . taking @xmath275 at @xmath276 , @xmath277 where we have purposefully left @xmath278 in its implicit form to provide a framework for the more mathematically complicated ii - m transition analysis later in this section . as one adjusts the gate voltage , @xmath279 passes through @xmath95 at @xmath280 , such that near the transition @xmath281 . to determine the critical behavior , we can therefore expand @xmath279 around this value . finite temperature @xmath282 can be taken into consideration by cutting off the renormalization group flow @xmath283 at @xmath284 . taking @xmath285 at an arbitrary ratio , the previous equation can be rewritten @xmath286 or explicitly inverted @xmath287\ ] ] where @xmath288 such that @xmath289 . with dashed lines overlaid to indicate possible voltage curves . as the gate voltage @xmath135 winds along a voltage curve , whose exact curvature is dictated by experimental specifics , it passes directly from the ii to the cc region along a curve like ( a ) or indirectly , passing along the way through an intermediate m phase along a curve like ( b ) . at zero temperature , the left - to - right conductance @xmath14 will therefore undergo a direct transition from @xmath290 along ( a ) or one with an intermediate step up to @xmath11 along ( b ) . this behavior motivates us to search for the finite - temperature scaling of these conductance transitions for @xmath291 ( a ) , or for @xmath292 and @xmath293 ( b).,width=336 ] @xmath294 is the universal critical exponent for the ii - cc quantum phase transition . figure [ figtemp ] shows @xmath295 at finite temperatures , noting that it collapses onto a step function at @xmath296 and that it has a crossover point pinned at @xmath297 for all values of interaction , making it identical to the t - r scaling function for weak interactions in the related quantum spin hall problem @xcite . when @xmath239 and the only transitions are directly from ii - cc , @xmath298 , where @xmath299 is the t - r critical exponent in ref . . this factor of @xmath96 accounts for the fact that even though here only a single linear combination of indexes is being pinched off , the diagrams which contribute to the flow equations still live in a matrix space which is four times as large as that of the comparable quantum spin hall problem . obtaining the critical exponent and scaling function for the ii - m quantum phase transition is procedurally identical . as an example , we will choose the bottom right transition point in fig . [ figzerot ] for our calculation , though that point is restricted to be the same as the one characterizing the finite @xmath190 ii - m transition by band - index - exchange symmetry . the quantum critical point is located at @xmath300 , @xmath301 where @xmath302 first and foremost , we can note that when @xmath303 , @xmath239 and there is no more available phase space ( at quadratic order ) for the ii - m transition to exist . near this case , the m phase will exist in a vanishing area of phase space and most transitions will be controlled by the central quantum critical point in figure [ figrg ] . however , in appendix [ appendix : stability ] we have only analytically calculated the stability of the ii - m quantum critical point to linear order in @xmath242 , whereas our analysis of the critical behavior of the ii - m transition is up to @xmath304 . we believe it reasonable to assume that this stability extends up to quadratic order in the interactions such that these quantum critical points still describe the relevant physical transitions in this problem . describes the direct ii - cc quantum phase transition and @xmath305 describes the ii - m transition , plotted in panels ( a ) and ( b ) respectively . here , we have plotted using @xmath306 , @xmath307 , such that @xmath308 . the curves are plotted for increasing temperature , with the red , orange , green , and blue curves representing @xmath309 , @xmath310 , @xmath311 , and @xmath312 respectively in equations ( [ gascaling ] ) and ( [ gbscaling ] ) . note that the crossover value of @xmath295 is fixed to be @xmath313 , whereas the crossover value for @xmath305 is instead at @xmath314 , where @xmath315 varies from @xmath95 to @xmath7 continuously as a function of interaction strength.,width=326 ] for the ii - m transition , the conductance jumps from @xmath316 , with here the @xmath189 axis being the only relevant direction . for this transition then , @xmath317 . expressing the conductance in its scaling form @xmath318 where again @xmath319 is the external gate voltage , @xmath39 is a non - universal constant that may certainly differ from the @xmath39 in the ii - cc transition , and the subscript @xmath16 denotes that transition between the ii - m phases . taking @xmath300 , the conductance transition from ii - m is characterized by the flow of a single parameter , @xmath320 we can determine the crossover scaling function by integrating this equation . taking @xmath321 at @xmath276 , @xmath322 as before , we can cut off the renormalization group flow at finite temperature @xmath323 and note that the gate voltage @xmath324 when @xmath325 passes through @xmath314 , such that near the transition @xmath326 . again taking @xmath285 at an arbitrary ratio , we can finally arrive at an implicit equation for @xmath327 , @xmath328 where @xmath329 such that @xmath330 . we can note that this equation reduces to eq . ( [ gascaling ] ) for @xmath331 . that case represents @xmath332 , for which @xmath189 and @xmath190 act as independent copies of the quantum spin hall problem in ref . , but in a higher - dimensional space . therefore , @xmath333 is the universal critical exponent for the ii - m transition . confirming the relationship to the qsh point contact problem , setting @xmath331 gives @xmath334 . figure [ figtemp ] shows @xmath305 as a function of @xmath319 at different temperatures , noting that at zero temperature it is also a step function , indistinguishable from @xmath295 , but that at finite temperature it is defined by a crossover value of @xmath314 which in general differs from that of @xmath295 ( fig . [ figscaling ] ) . taking advantage of the duality between the ii and cc phases , we can relate the remaining conductance crossover function , one which characterizes the m - cc phase transition , to @xmath305 . we can write the conductance in its scaling form @xmath335 where @xmath39 is yet another non - universal constant and @xmath336 denotes that m - cc transition . by pinch - off symmetry , we know that @xmath98 for the cc@xmath337 m transition has to be equivalent to @xmath14 for the ii@xmath337 m transition , therefore for the m@xmath337cc transition , @xmath338 and utilizing eq . ( [ 2term ] ) , @xmath339 such that we finally deduce @xmath340 where @xmath341 and @xmath342 for weak interactions . in this paper , we have computed the conductance signatures of the four - terminal intersection of two bilayer graphene domain walls . these domain walls can be induced by the presence of a perpendicular electric field and a change in either electric field direction or interlayer stacking . when valley - index is conserved , the domain walls are luttinger liquids described with two non - trivial interaction parameters @xmath46 . the junction is analogous to a point contact and can be analyzed naturally using the language of quantum point contacts . as with a quantum spin hall point contact , the physics of the junction is best understood in terms of reduced , two - terminal conductances . when interactions are strongly attractive ( @xmath343 ) or strongly repulsive ( @xmath344 ) , the left - to - right conductance can be strictly dominated by nonzero charge and valley conductances respectively . for weaker interactions ( @xmath345 ) , both left - to - right conductances are nonzero and there exist several stable phases characterized by single - electron tunneling . transitions between these phases are governed at low temperatures by universal scaling functions and critical exponents , which differ from those in the single - band qsh case and are functions of the two luttinger parameters . we now briefly discuss the task of experimentally measuring the physics in this paper . first and foremost , the existence of a single domain wall in bilayer graphene requires prohibiting scattering between valleys @xmath1 and @xmath2 . valley - index - breaking perturbations are strongly relevant and will significantly change the physics in both isolated domain walls and for the junction structures at their intersections . to this end , short - range disorder must be kept smooth on the scale of the lattice . recent experiments have shown promising results in this direction , with fabricated samples showing single domain wall conductances up to nearly the quantized clean limit of @xmath11 @xcite . beyond these results , one should then attempt to verify the luttinger liquid physics at a single domain wall by measuring the tunneling conductance at several low temperatures and its collapse onto a universal scaling function with critical exponent @xmath12 . in junction structures , the conservation of valley index can be confirmed by measuring the quantization of the reduced conductance element @xmath265 . creating a multi - terminal junction like the ones we describe poses several fabrication and analysis difficulties which must be overcome to measure the point - contact physics in this paper . forming a four - terminal junction of electric - field - induced domain walls requires patterning leads on the top and bottom of each of the four bulk regions , as well as a gate on the junction to control the weak - interaction pinch - off transition . forming a junction from layer - stacking domain walls requires patterning fewer leads , but as clearly demonstrated in the samples in ref . , the three - fold symmetry of the underlying graphene lattice restricts intersections of these domain walls to be six - terminal structures . conductance transitions in these six - terminal structures can be calculated and analyzed using the framework established in this paper , though the task will be algebraically more intensive . tuning the two luttinger parameters @xmath46 can be accomplished through turning a combination of experimental knobs . the strength of the overall effective coulomb interaction can be altered for the domain wall states by tuning their widths with the strength of the perpendicular electric field . these changes will be mainly reflected in @xmath229 , as it contains long - range contributions from the overall coulomb interaction . the other luttinger parameter , however , can only be adjusted by tuning short - range interactions . this can be accomplished by testing samples on a variety of substrates . working in order of increasing dielectric strength , one can work through a suspended sample , a silicon dioxide substrate , or a boron nitride substrate to tune down @xmath228 . several simplifications and assumptions in this paper may not be exactly present under experimental conditions . the assumption that all domain walls in the sample have the same values of the luttinger parameters requires that the perpendicular electric field strength be globally uniform in magnitude in the bulk regions and change in direction similarly smoothly across electric - field - induced domain walls . it also requires that there are no strong local variations in the dielectric strength and coupling of the underlying substrate . the physics of junctions may be significantly altered if these conditions are not realized experimentally ; we have not investigated point contacts at the intersections of two domain walls with differing interaction strengths . we also only calculated universal scaling functions to leading order in the interactions about the noninteracting point @xmath121 . when interactions become stronger and single - electron tunneling is no longer the least irrelevant operator , the critical exponents @xmath18 in the left - to - right conductance transitions may change significantly in their dependences on the interaction strengths , as they do in ref . . finally , we assumed that the fermi energy was exactly at the particle - hole - symmetric point such that @xmath346 . in practice , it will be quite difficult to exactly tune the fermi energy to this point , and so for most experimental realizations , @xmath187 exchange symmetry in the variables will be broken . for our analysis , the effects of this can be realized by replacing the equal band - index - exchange symmetry which we exploited with one which flips band index and rescales variables by @xmath347 . this will result in changes to the scaling dimension calculations in section [ sec : fourterminal ] and a relaxation of mirror symmetry about @xmath234 in figure [ figrg ] . this work was supported by a simons investigator award from the simons foundation to charles kane . fan zhang was supported by ut dallas research enhancement funds . in this appendix , we confirm analytically the stability of the two classes of quantum critical points in fig . [ figrg ] which control the transitions between single - electron - tunneling phases . here , we begin with the general flow equation for elements of the s - matrix ( eq . ( [ sflow ] ) ) , leaving free the @xmath181 and @xmath182 variables in the full s - matrix formulation from [ sec : noninteracting ] . as the complexity of this calculation greatly grows with each power of @xmath242 taken into consideration , we will only here carry out our stability analysis to quadratic order in @xmath348 and linear order in @xmath242 . our methodology can be used to analytically calculate higher - order terms , but we believe it reasonable to truncate the calculation at this order and that the stability should carry over to the @xmath304 calculation used to produce the phase diagrams and scaling functions in [ sec : interactingrg ] . additionally , numerical estimates find @xmath349 for a large range of system parameters @xcite , therefore the limitation to terms of order @xmath350 is physically motivated . we can exploit the matrix structure of the diagrammatics by taking several traces of s - matrix products . derivatives of these traces exploit the matrix structure of the flow equation , effectively closing the external legs of the diagrams in fig . [ figdiagram ] into additional loops . we will follow through this calculation completely for a single trace ( @xmath351 $ ] ) and then show how taking linear combinations of these traces results in a flow equation for @xmath352 . all remaining @xmath353 can be obtained by exploiting cyclic reindexing and pinch - off symmetries . to begin this process , let us rewrite s - matrix elements in the @xmath170 band space using a simplified version of eqs . ( [ smatrix ] ) , ( [ umatrix ] ) : @xmath354 @xmath355 where @xmath356 is a collection of scalars in the band index space with @xmath30 , @xmath357 and all matrix operations only involve the @xmath114 indexes . transmission probabilities are normalized for each band such that @xmath358 and the labeling @xmath359 on the phases corresponds to : @xmath360 we ll begin by calculating three traces : @xmath361 & = & |d_{ij}^{0}|^{2 } + |d_{ij}^{\pi}|^{2 } \nonumber \\ \operatorname{tr}[s_{ij}^{\dagger}\sigma_{z}s_{ij } ] & = & \left(|d_{ij}^{0}|^{2 } - |d_{ij}^{\pi}|^{2}\right)\cos{\theta_{i } } \nonumber \\ \operatorname{tr}[s_{ij}^{\dagger}\sigma_{x}s_{ij } ] & = & \left(|d_{ij}^{0}|^{2 } - |d_{ij}^{\pi}|^{2}\right)\sin{\theta_{i}}. \label{2s}\end{aligned}\ ] ] we can then differentiate and take weighted linear combinations of these traces to obtain explicit flow equations for @xmath181 . specializing to @xmath362 , @xmath363 \nonumber \\ & - & ( \sin{\theta_{1}})\frac{d}{dl}\operatorname{tr}[s_{12}^{\dagger}\sigma_{z}s_{12 } ] \end{aligned}\ ] ] where @xmath364 . now , we can examine , in detail , the process of using the diagrammatics ( eq . ( [ sflow ] ) ) to calculate one of these derivatives , namely @xmath365 $ ] . all other trace derivatives , while varying in signs and specifics , follow procedurally from this example . first , by using the cyclic index definition of the trace , we can see that the derivative of the trace is equal to the trace of the chain rule derivative : @xmath366 & = & \operatorname{tr}[\frac{ds^{\dagger}_{12}}{dl}\sigma_{z}s_{12 } ] + \operatorname{tr}[s^{\dagger}_{12}\sigma_{z}\frac{ds_{12}}{dl } ] \nonumber \\ & = & \operatorname{tr}[s^{\dagger}_{12}\sigma_{z}\frac{ds_{12}}{dl } ] + c.c.\end{aligned}\ ] ] where for the second equality we exploited the cyclic nature of the trace to reduce this step to the calculation of one , albeit large , trace . in this case , the trace of the derivative is real , but in the next section where we calculate @xmath367 , it will not be and the addition of the complex conjugate can not be overlooked . from here , the calculation amounts to taking the traces of terms which contain products of two or four s - matrices . while the products of two , in the form of eq . ( [ 2s ] ) , can be calculated by rote algebra without much difficulty , the terms with four s - matrices require a bit more manipulation . we calculate the traces of products of four s - matrices by both recognizing a pattern in the assignment of signs to the products of @xmath368 and with a careful treatment of commutivity issues . consider first the simplest example , @xmath369 $ ] . without any additional pauli matrices , the @xmath174 rotations all cancel out pairwise , resulting in : @xmath370 = \operatorname{tr}[(a_{l2}^{*}\mathbb{1 } + b_{l2}^{*}\sigma_{z})(a_{lk}\mathbb{1}+b_{lk}\sigma_{z } ) \nonumber \\ ( a_{1k}^{*}\mathbb{1 } + b_{1k}^{*}\sigma_{z})(a_{12}\mathbb{1 } + b_{12}\sigma_{z})].\ \end{aligned}\ ] ] one might be concerned that converting this to a useful form , one with @xmath371 where s - matrix elements can just be read off , would be a daunting and terrible task . however , converting to @xmath371 basis is equivalent to summing over all of the ways to choose @xmath122 and @xmath123 signs for the cross terms , and so for every single one of these four s - matrix traces , _ all of the cross terms cancel_. our trace is reduced to the quite simple form : @xmath372=\sum_{a=0,\pi}(d_{l2}^{a})^{*}d_{lk}^{a}(d_{1k}^{a})^{*}d_{12}^{a}. \label{simple4trace}\ ] ] worth noting is that this picture is significantly complicated by the addition of pauli matrices between s - matrix factors , due to the fact that @xmath373\neq 0 $ ] . while the complexity does nt significantly increase for the addition of a single pauli matrix , as it can be absorbed into the definition of @xmath374 , it does for two or more pauli matrices . this can be seen at the level of the two - s - matrix trace , where commutivity issues lead to the addition of a second term : @xmath375&=&\cos{\theta_{i}}\cos{\theta_{j}}\left[|d_{ij}^{0}|^{2 } + |d_{ij}^{\pi}|^{2}\right ] \nonumber \\ & + & \sin{\theta_{i}}\sin{\theta_{j}}\left[d_{ij}^{0}(d_{ij}^{\pi})^{*}+(d_{ij}^{0})^{*}d_{ij}^{\pi}\right].\nonumber \\ \end{aligned}\ ] ] for the four - s - matrix traces , the weighting of the cross terms is altered by the anticommutivity of the pauli matrices and while only two terms remain for each trigonometric function of @xmath181 , they are different than the simple form of eq . ( [ simple4trace ] ) and contain possible terms which mix @xmath9 and @xmath19 . as the number of pauli matrices inserted into these traces increases linearly with the power of @xmath242 to which we expand , we have chosen for the sake of simplicity and clarity to expand only to linear order in @xmath242 . though our analysis in section [ section : qcps ] continues to @xmath304 , we believe that the stability deduced here should carry over to higher order terms . with our calculation machinery established , we can produce a flow equation for the @xmath187 mixing at lead @xmath95 : @xmath376 \nonumber \\ & + & \sin{2\theta_{2}}\cos\left(2\phi_{21}\right)\sqrt{t_{0}t_{\pi}}[2-t_{0}-t_{\pi } ] \nonumber \\ & + & 2\sin{2\theta_{3}}\cos\left(2\phi_{31}\right)\sqrt{t_{0}(1-t_{0})t_{\pi}(1-t_{\pi } ) } \nonumber \\ & + & \sin{2\theta_{4}}\cos\left(2\phi_{41}\right)\sqrt{(1-t_{0})(1-t_{\pi})}[t_{0}+t_{\pi}]\bigg\}.\nonumber \\\end{aligned}\ ] ] flow equations for the remaining three mixing angles can be generated by exploiting underlying symmetries of the problem . the set of nine independent variables which characterizes the s - matrix obeys three symmetries . two `` pinch - off '' symmetries exist ; the duality between the fully closed ( ii ) and fully open ( cc ) single - electron phases implies that the system of flow equations is invariant with respect to the exchange @xmath232 and _ either _ the exchange of lead indexes @xmath377 _ or _ @xmath378 . the third symmetry is a cyclic relabeling of the lead indexes as well as an exchange of the definitions of pinched off and open , due to the system s invariance under properly - treated ( with respect to valley ) @xmath110 rotations . therefore the system is also invariant under the exchange @xmath232 and @xmath379 , @xmath380,@xmath381 , and @xmath382 . independent calculations of @xmath353 confirm these properties . we can see immediately that for @xmath239 , all @xmath181 are marginal . in this case , the system has u(2 ) symmetry at each lead and all band index rotations can be gauged out . the dependence of @xmath353 on @xmath383 can also be suppressed by tuning the scattering phase @xmath182 closer to @xmath384 , which amounts to having a @xmath385 scattering phase difference between the @xmath9 and @xmath19 bands . calculating the flow equations for scattering phases @xmath182 requires , conversely , tracing over open paths in diagrams , which allows phase to accumulate throughout the summation instead of being canceled out pairwise as frequently occurred in the calculation of @xmath353 . specializing for the moment towards obtaining @xmath386 , we can note the following : @xmath387 & = & ( t_{0}e^{i\phi_{a } } + t_{\pi}e^{-i\phi_{a}})\cos\left(\frac{\theta_{1}-\theta_{2}}{2}\right ) \nonumber \\ where the correspondence between the subscripts @xmath388 and the lead indexes @xmath159 comes from eq . ( [ phidef ] ) . this can be differentiated , and , with a considerable amount of algebra , used to obtain first order flow equations for the scattering phases . while the specifics of this calculation differ from those of obtaining the @xmath353 , the key point about the summation over @xmath122 and @xmath123 possibilities when converting to the @xmath371 basis remains for both the one- and three - s - matrix products here , again greatly simplifying the algebra for the required trace calculations . utilizing this fact , we obtained an explicit flow equation for scattering phase @xmath389 : @xmath390 \nonumber \\ & -&\sin{2\theta_{4}}\sqrt{(1-t_{0})(1-t{\pi})}\left[(t_{0}+t_{\pi})\sin\left(2\phi_{b}\right ) - 2\sqrt{t_{0}t_{\pi}}\sin\left(2(\phi_{a}-\phi_{b})\right)\right]\bigg\}.\end{aligned}\ ] ] the flow equations for @xmath391 can be obtained by exploiting pinch - off and cyclic symmetries as well as the redundancy of @xmath392 ( eq . ( [ phidef ] ) ) . as with the @xmath181 , one can observe that for the @xmath393 symmetric case of @xmath239 , all @xmath182 are also marginal and can be gauged away . one can also observe that @xmath394 when @xmath395 , as in that case there is no relative interband scattering between leads @xmath176 and @xmath177 and @xmath182 can be gauged out . only two unique stability calculations are required , as the four critical points on the boundary of the square are related by pinch - off and band - index - exchange symmetries . we can now utilize the flow equations for @xmath181 and , to a lesser extent , @xmath182 to determine the stability of the @xmath396 surface quantum critical points in fig . [ figrg ] . we will begin by considering the @xmath240 central quantum critical point , which controls the direct transition between the fully pinched - off ( ii ) and fully - open ( cc ) phases . expanding the @xmath181 to linear order : @xmath397 @xmath398 which gives two lyapunov exponents of zero and two of @xmath399 . both @xmath65 eigenvalues are either zero or negative for all choices of the independent @xmath400 . given the high symmetry of this central quantum critical point , it is unlikely that the @xmath401 marginal directions correspond to instabilities at higher orders . the four external critical points which control the transitions to and from the mixed ( m ) phase can be handled similarly , but with an attention to the expansions used to arrive at this analysis . as our model is invariant under global exchange @xmath187 and the pinch - off symmetries , we perform stability analysis on just one of the four critical points and relate the rest by symmetry . choosing the critical point at @xmath402 , we can again construct a stability matrix : @xmath403 @xmath404 acknowledging that , as we have expanded to linear order in @xmath242 , @xmath405 , all of the off - diagonal terms become infinitesimal and we are left with : @xmath406 which clearly has strictly negative lyapunov exponents up to this order of expansion . the flow diagram in fig . [ figquad ] , obtained by perturbation theory to quadratic order in the interactions , suitably describes the transport physics of single - electron - tunneling phases for most ranges of infinitesimal interactions @xmath348 and @xmath242 . however , when @xmath239 , eqs . ( [ floweq ] ) become symmetric under arbitrary u(2 ) transformations in band index space and describe a fixed _ line _ @xmath407 . in this appendix , we use bosonization at the @xmath408 , @xmath300 `` m - phase '' fixed point to determine whether this fixed line is a genuine physical phenomenon or an artifact of our @xmath409 perturbation theory . the first section of this appendix details the derivation of a euclidean action about an m fixed point , working from the action for the fully - open cc phase . the second section utilizes that action to calculate flow equations beyond quadratic order in the vicinity of the m phase , resolving the fixed line to additional fixed points for weak - to - moderate interactions . to examine the higher - order behavior of tunneling processes about the @xmath246 corner fixed point in eq . ( [ floweq ] ) , we have to first obtain a theory for that fixed point in terms of our existing theory for the @xmath411 , fully - open fixed point . as we will be focusing on the fate of the quadratic - order fixed line at @xmath250 , we will for the purposes of this calculation specialize to @xmath412 , @xmath250 . integrating out the @xmath413 sectors in eq . ( [ spinfulll ] ) , we can propose as a starting point the euclidean action about the cc fixed point : these processes will generally be present within the vicinity of the @xmath419 fixed point . both of these processes are just single - electron tunneling , so as described in [ sec : fourterminal ] , they are marginal or irrelevant ( @xmath420 $ ] ) for all values of @xmath421 , such that we may arbitrarily increase the coupling strength @xmath422 without new , additional tunneling processes becoming relevant . respecting time - reversal , valley - index , and band - index symmetries , @xmath423 are restricted to live in the plane of fig . [ figrg ] and therefore turning up @xmath422 represents motion away from the cc fixed point along the @xmath424 axes respectively . for this reason , we need to return @xmath425 to the @xmath426 basis so that we can generate an m phase action by taking the tunneling strength @xmath427 and `` pinch off '' just the @xmath19 band . additionally , we will find it easier to work in a basis for which the tunneling operators @xmath423 are just sums of cosines . from our bosonization work in [ sec : model ] and utilized earlier in this appendix , we recall that @xmath428 . we will also need to transform valley indexes utilizing the property that @xmath429 where @xmath30 . combining these definitions , we create the unitary change - of - basis to drive towards the @xmath443 fixed point , we need only take @xmath427 , as we have already established @xmath444 and @xmath445 as the quantities which push back along the @xmath189 and @xmath190 axes respectively . to massage something useful out of this strong - tunneling limit , we can utilize an extreme limit of the villain approximation , following procedurally a longer formulation of the kane - fisher problem @xcite . note that the partition function is the product of the contribution from @xmath425 and a term of the form @xmath446 . as @xmath427 , the entire partition function is zeroed out except when @xmath447 is a minimum . acknowledging this , one can make the substitution of @xmath448 where @xmath449 is now a discrete step in time . the partition function can then be integrated over @xmath79 once we complete the square to give a new gaussian effective action in terms of the integer @xmath449 . now , we can define @xmath450 , such that in the frequency domain , our bare action has _ the same form and an inverse luttinger parameter_. hopping events between the minima of @xmath447 are instantons whose term in the action takes the form @xmath451 , the same as if we had defined our electron operators as exponentials of @xmath452 fields and examined tunneling about the bare action . this equivalence between strong electron tunneling in @xmath79 and weak quasiparticle backscattering in @xmath452 is a key feature of the kane - fisher problem for single - impurity scattering in luttinger liquids . its physical implications , as well as a much more detailed derivation of it , can be found in ref . . for four - terminal quantum point contacts , however , this equivalence manifests itself as a relationship between strong left - to - right electron tunneling and weak top - to - bottom electron backscattering @xcite . therefore , our ultimate goal is to arrive at @xmath453 operators which correspond to the @xmath454 single - particle tunneling processes in fig . [ figtv ] about an m point in fig . [ figrg ] , for which only the @xmath19 band is pinched off . as addressed in eq . ( [ ccaction ] ) , the full action contains contributions from @xmath425 as well as perturbative tunneling terms . to push towards the corner m phase , we can increase @xmath445 greatly until it becomes valid to replace that part of the partition function with discrete delta functions in @xmath458 : such that @xmath461 . @xmath190 is an instanton tunneling term which enforces the condition that @xmath462 is an integer . the factor of @xmath7 in front of it is a normalization requirement from the @xmath463 change of basis . as described earlier , we can recognize the @xmath462 as discrete steps in time of a new field : after a good bit of algebra , we can complete the square twice here and integrate out the @xmath458 , leaving us with an intermediate action @xmath468 about the @xmath408 , @xmath300 m - phase stable fixed point : in which the sign on the bottom right block of @xmath473 is owed to one of the sets of bosonic fields being evaluated at @xmath474 and the other at @xmath475 . the tunneling term @xmath476 for the @xmath452 fields represents weak , left - to - right tunneling of electrons with band index @xmath19 ; making it very large would return the system back to the cc corner fixed point . where the factor of @xmath7 again comes from the spin - index change of basis and ensures that diagonal correlators like @xmath478 are appropriately marginal . the calculations in the following section are also only appropriate provided that the single - particle , _ off - diagonal _ correlation functions are zero . this requirement in enforced for the @xmath79 fields for @xmath479 and by @xmath480 for the @xmath452 variables . however , the range @xmath481 is satisfied by all physical values of @xmath421 . therefore , the only bounds on the validity of the @xmath468 theory are the regions in fig . [ figphase ] for which many - body processes become relevant . in this section , we will , using the @xmath468 theory established in the previous section , develop quartic - order flow equations for @xmath444 and @xmath476 for @xmath483 . after establishing the flow in those variables , we will relate @xmath444 and @xmath476 to @xmath424 in eq . ( [ floweq ] ) . expanding @xmath484 , we will develop the cubic - order corrections to fig . [ figquad ] close to the m - phase fixed points and obtain the schematic phase diagram fig . [ figrg ] . there are two sets of correlation functions for each coupling coefficient that we must calculate and rescale to obtain flow equations , as indicated by the two non - zero , off - diagonal elements for each operator and spin in @xmath473 . for now , we will work just with the @xmath444 renormalization and then use band - index - exchange and pinch - off symmetries to relate them to the flow of @xmath476 . @xmath486\rangle _ { > } - \langle e^{i\theta_{0\uparrow k}(\tau)}\rangle_{>}\langle t_{\tau}\left[e^{i\theta_{0\sigma k'}(\tau_{1})}e^{-i\theta_{0\sigma k'}(\tau_{2})}\right]\rangle _ { > } \nonumber \\ & + & v_{0}t_{\pi}^{2}\sum_{\sigma=\uparrow,\downarrow}\int_{-\infty}^{\infty}\ d\tau_{1}d\tau_{2}\langle t_{\tau}\left[e^{i\theta_{0\uparrow k}(\tau)}e^{i\varphi_{\pi\sigma k'}(\tau_{1})}e^{-i\varphi_{\pi\sigma k'}(\tau_{2})}\right]\rangle _ { > } - \langle e^{i\theta_{0\uparrow k}(\tau)}\rangle_{>}\langle t_{\tau}\left[e^{i\varphi_{\pi\sigma k'}(\tau_{1})}e^{-i\varphi_{\pi\sigma k'}(\tau_{2})}\right]\rangle_{>}. \nonumber \\\end{aligned}\ ] ] @xmath487 \nonumber \\ & + & 2v_{0}t_{\pi}^{2}\int_{-\infty}^{\infty}\ d\tau_{1}d\tau_{2}\ \frac{1}{\tau_{12}^{2}}\left[\theta(\tau\not\in(\tau_{1},\tau_{2 } ) ) + e^{\pi i\alpha}\theta(\tau_{1}<\tau<\tau_{2 } ) + e^{-\pi i\alpha}\theta(\tau_{2}<\tau<\tau_{1 } ) -1\right]\end{aligned}\ ] ] where @xmath488 and @xmath489 is the heaviside step function expressed in conditional notation . these integrals can be performed analytically and , after a bit of work , we arrive at an explicit equation for @xmath490 : now , to answer our initial question , we will examine the consequences of our new flow equations near the vicinity of the possible fixed line in fig . [ figquad]b . expanding @xmath484 , @xmath496 such that eq . ( [ eqtv ] ) reduces to the @xmath239 limit of eq . ( [ floweq ] ) under the substitution : when @xmath498 . we also know that the linear terms in @xmath499 must agree with any local expansion in the @xmath253 plane , such that we can restore flow to linear order in @xmath242 near the m - phase by extracting the leading term in eq . ( [ floweq ] ) : where we have used the same approximation @xmath498 . combining this with an expansion of eq . ( [ eqtv ] ) to quartic order in @xmath348 , we obtain , finally , higher - order flow equations about an m - phase fixed point : @xmath501 - 4(1-t_{0})t_{\pi}\left[\epsilon_{+}^{2 } - \frac{\pi^{2}\epsilon_{+}^{4}}{48}\right ] \nonumber \\ \frac{dt_{\pi}}{dl } & = & -8\epsilon_{+}\epsilon_{-}t_{\pi } + 4t_{\pi}^{2}\left[\epsilon_{+}^{2 } + \frac{\pi^{2}\epsilon_{+}^{4}}{48}\right ] - 4(1-t_{0})t_{\pi}\left[\epsilon_{+}^{2 } - \frac{\pi^{2}\epsilon_{+}^{4}}{48}\right].\end{aligned}\ ] ] equation ( [ quartflow ] ) contains a few points of interest . first and foremost , it reflects the pinch - off symmetry in its invariance under the exchange @xmath502 . it also contains fixed points at @xmath503 and at @xmath504 , in agreement with the small @xmath236 limit of fig . [ figquad ] . we specifically normalized our correlations about @xmath468 to fix this agreement , such that we could locate at quartic order in @xmath348 any local quantum critical points which control the transition between the cc / ii and m phases . relevant flow lines from this point head off towards the central @xmath240 critical point , _ even in the @xmath239 case when this point converges with the other critical points at the corner_. therefore , this point stands as a demonstration that the fixed line in fig . [ figquad]b is an artifact of @xmath507 perturbation theory , at least in the vicinity of the @xmath246 m point . it is most likely that this flow away from the m point continues , to lowest order , all the way to the central quantum critical point . this infers that the phase diagram of this system for weak - to - moderate interactions is best described by the schematic fig . [ figrg ] . j. velasco jr . , l. jing , w. bao , y. lee , p. kratz , v. aji , m. bockrath , c. n. lau , c. varma , r. stillwell , d. smirnov , f. zhang , j. jung , and a. h. macdonald , nature nanotechnology * 7 * , 156 ( 2012 ) . the related discussions were provided in the supplementary materials .
bilayer graphene in a perpendicular electric field can host domain walls between regions of reversed field direction or interlayer stacking . the gapless modes propagating along these domain walls , while not strictly topological , nevertheless have interesting physical properties , including valley - momentum locking . a junction where two domain walls intersect forms the analogue of a quantum point contact . we study theoretically the critical behavior of this junction near the pinch - off transition , which is controlled by two separate classes of non - trivial quantum critical points . for strong interactions , the junction can host phases of unique charge and valley conductances . for weaker interactions , the low - temperature charge conductance can undergo one of two possible quantum phase transitions , each characterized by a specific critical exponent and a collapse to a universal scaling function , which we compute .
let @xmath2 be a tame symplectic manifold ( * ? ? ? * definition 4.1.1 ) . consider a pair of closed lagrangian submanifolds @xmath3 . in this paper we will concentrate on the equivalence relation of lagrangian cobordism between lagrangian submanifolds @xmath4 . * definition 2.1.1 ) an ( elementary ) lagrangian cobordism @xmath0 is a smooth cobordism @xmath5 between @xmath6 and @xmath7 admitting a lagrangian embedding @xmath8\times \mathbb{r})\times m , ( dx \wedge dy ) \oplus \omega)\ ] ] such that for some @xmath9 we have : @xmath10 @xmath11 \times \mathbb{r}\times m ) = ( 1-\epsilon , 1 ] \times \{1\ } \times l_{1}.\ ] ] the relation of lagrangian cobordism was introduced by arnold @xcite , @xcite . since then , many authors have studied this relation @xcite , @xcite,@xcite,@xcite,@xcite , @xcite . a recent result of biran and cornea ( * theorem 2.2.2 ) , states that if a pair of monotone lagrangian submanifolds @xmath4 , are related by a monotone lagrangian cobordism @xmath0 , then their associated floer homologies ( when defined ) are isomorphic , that is @xmath12 . a choice of orientation and spin structure on the cobordism @xmath0 allows us to extend this result to @xmath13-coefficients . thus , we can conclude that the @xmath13-floer homologies of the lagrangians @xmath4 coincide ( see the precise statement as theorem [ teorema ] ) . consider an orientable and spin exact lagrangian cobordism @xmath14 satisfying the topological condition : _ [ hyp ] ( 1 ) the morphisms @xmath15 and @xmath16 induced by the inclusion maps @xmath17 and @xmath18 are isomorphisms . _ for such a cobordism , theorem [ teorema ] implies that @xmath0 is an h - cobordism . by an h - cobordism we understand a cobordism @xmath0 where the inclusion map @xmath19 is a homotopy equivalence for @xmath20 . in summary , an orientable - exact - spin - lagrangian cobordism satisfying condition [ hyp ] is an h - cobor- + dism . in particular , for a simply connected exact lagrangian cobordism , the @xmath21-version of ( * ? ? ? * theorem 2.2.2 ) and the h - cobordism thoerem of smale , imply that the cobordism is a pseudo - isotopy i.e. it is diffeomorphic to the trivial cobordism @xmath22\times l_{0 } ; l_{0 } , l_{0})$ ] . this result appears in @xcite , for an exact lagrangian cobordism in a simply - connected cotangent bundle and the proof uses different techniques . in higher dimensions @xmath23 , the h - cobordisms are classified ( up to simple homotopy equivalence ) by their whitehead torsion . for these dimensions , the @xmath24-cobordism theorem states that an h - cobordism with trivial whitehead torsion is a pseudo - isotopy . it is natural to ask whether an exact lagrangian cobordism @xmath0 is a lagrangian pseudo - isotopy ( by a lagrangian pseudo - isotopy we understand a lagrangian cobordism diffeomorphic to the trivial cobordism @xmath22\times l_{0 } ; l_{0 } , l_{0})$ ] ) . given the rigidity of exact lagrangian submanifolds , and therefore of exact lagrangian cobordisms , biran and cornea proposed the following conjecture : any exact lagrangian cobordism is hamiltonian isotopic to a lagrangian suspension . a lagrangian suspension @xcite is a lagrangian pseudo - isotopy defined by the image of the map @xmath25\times l & \rightarrow [ 0,1 ] \times \mathbb{r } \times m\\ ( t , x)&\rightarrow(t , -h(t,\psi^{t}(x)),\psi^{t}(x))\end{aligned}\ ] ] where @xmath26}$ ] is a hamiltonian isotopy generated by some hamiltonian @xmath27 and @xmath28 is some lagrangian submanifold . a weaker form of the conjecture of biran - cornea is proven in the main theorem of this paper : let @xmath0 be an exact , orientable and spin lagrangian cobordism equipped with a choice of spin structure . assume @xmath29 . if the map @xmath30 induced by the inclusion @xmath31 is an isomorphism for @xmath32 , then @xmath0 is a lagrangian pseudo - isotopy . the paper is organized as follows . in section [ chapter floer homology ] we study the homology of the @xmath13-floer complex of a monotone lagrangian cobordism . the technical aspect of the section concerns the definition of the floer complex for lagrangian cobordism , and is treated following the work of biran - cornea @xcite . as a result , we obtain that a cobordism which satisfies the hypotheses of the main theorem is automatically an h - cobordism , and as such it has a well - defined whitehead torsion . in section [ chapter simple h - theory ] we briefly recall some definitions and known results on the whitehead torsion . in particular , we state the s - cobordism theorem @xcite , @xcite , @xcite , which will constitute a fundamental ingredient in the proof of the main theorem . in section [ chapitre whitehead ] we study the invariance ( under horizontal isotopies ) of the whitehead torsion of the @xmath13-floer complex associated to an exact lagrangian cobordism ( see theorem [ theorem invariance ] ) . this is , perhaps , the most technical part in the proof of the main theorem due to the bifurcation analysis involved . the proof of theorem [ theorem invariance ] is based on the work of m. sullivan @xcite and lee @xcite . we apply the technique of stabilization of sullivan , and a gluing theorem for a degenerate floer strip , proved by lee for the floer theory of hamiltonian orbits . it follows that if a cobordism satisfies the hypotheses of the main theorem , then its torsion agrees with the torsion of the floer complex . we then apply the displaceability of the cobordism , which yields that the whitehead torsion of the cobordism vanishes . hence , we conclude with an application of the s - cobordism theorem , that the cobordism in question , is a lagrangian pseudo - isotopy . this concludes the proof of the main theorem , modulo orientation issues . finally , in section [ chapter orientations ] we address the orientation issues which arose in sections [ chapter floer homology ] and [ chapitre whitehead ] . this technical part is based on the work of fooo @xcite and lee @xcite . * acknowledgments * this paper is based on my ph.d thesis being carried out under the guidance of prof . octav cornea in universit de montral . this work would have not been possible without all the ideas and support of my great adviser , i am deeply grateful to him . i would like to thank jean - franois barraud , baptiste chantraine and franois lalonde for their interest in this work and helpful conversations . i am grateful to michael sullivan for his comments and corrections on my thesis , from which this paper has been extracted . one of the main tools used to study lagrangian submanifolds is floer homology . in this section we study floer homology with twisted coefficients , associated to a monotone lagrangian cobordism @xmath0 . in this section we show that the floer homology with twisted coefficients is invariant under monotone lagrangian cobordism , this is the twisted version of ( * ? ? ? * theorem 2.2.2 ) . moreover , when the lagrangian cobordism is exact , under an additional topological condition the cobordism itself is an h - cobordism . to a lagrangian cobordism @xmath0 we can associate a lagrangian with cylindrical ends @xmath33 ( see definition 2.2 ) . let @xmath34)$ ] denote the floer homology with @xmath13-coefficients of @xmath33 , where @xmath35 $ ] denotes the path component of a function @xmath36 defining a vertical perturbation on the ends of @xmath33 for @xmath37 . the main theorem of this section is the following : [ teorema ] let @xmath0 be an orientable and spin , monotone lagrangian cobordism equipped with a choice of spin structure . then , @xmath38 ) \cong 0 \cong hf(\overline{w},\overline{w};\mathbb{z}(\pi_{1}w),[f_{l_{1 } } ] ) \text { and,}\ ] ] @xmath39)\cong hf(l_{1};\mathbb{z}(\pi_{1}w)).\ ] ] as main consequences we will show the following corollaries : [ col1 ] if @xmath14 is an orientable and spin , exact lagrangian cobordism and @xmath40 is an isomorphism for @xmath41 , then @xmath0 is an h - cobordism . [ col2 ] if @xmath0 is an orientable and spin , exact lagrangian cobordism with @xmath42 and @xmath43 ( @xmath41 ) , @xmath5 are simply connected , then @xmath0 is a lagrangian pseudo - isotopy . corollary [ col2 ] has been shown in @xcite for the cotangent bundle using different techniques involving previous work of abouzaid @xcite and kragh @xcite . to extend the result of corollary [ col2 ] to the non - simply connected setting , it is necessary the study of the whitehead torsion , in order to apply the s - cobordism theorem . given a lagrangian @xmath45 there are two canonical homomorphisms , the symplectic area @xmath46 & \mapsto \omega(u ) : = \int_{d^{2 } } u^{*}\omega\end{aligned}\ ] ] and the maslov index : @xmath47 the _ minimal maslov number _ of @xmath48 is the integer @xmath49 defined by @xmath50 a lagrangian submanifold @xmath51 is monotone if there exist a constant @xmath52 , such that,@xmath53 for all @xmath54 and @xmath55 . we call @xmath56 the monotonicity constant . the lagrangian cobordisms @xmath14 , are orientable , monotone ( @xmath57 is a monotone lagrangian ) and spin and we assume that the choice of spin structure on @xmath5 , is such that when we restricted it to @xmath43 it coincides with the chosen spin structures on @xmath43 for @xmath58 . a fundamental assumption to have a well defined floer complex is that the minimal maslov number of any lagrangian @xmath48 , @xmath49 is strictly larger than two * @xmath59*. we will explain this latter . let @xmath60 be a pair of lagrangian submanifolds . in this subsection we recall the construction of the floer complex with @xmath44-coefficients associated to the pair @xmath61 . we also show the invariance under hamiltonian perturbation of the given homology . we follow the exposition in @xcite . the floer complex with twisted coefficients has been studied before in @xcite and a similar version called lifted floer complex is defined in @xcite . _ ring of coefficients and set of generators . _ consider the following rings : _ the integral group ring _ @xmath62 of the group @xmath63 ( in our case @xmath64 ) , is the set of finite formal sums : @xmath65 with the natural addition and multiplication . _ the universal novikov ring _ is denoted by @xmath66 ; @xmath67 with the addition and multiplication of power series . the set of generators is a subset of the space of paths in @xmath68 connecting @xmath6 with @xmath7 , @xmath69,m ) \hspace{0.2cm}|\hspace{0.2 cm } \gamma(0 ) \in l_{0 } , \gamma(1)\in l_{1 } \}.\ ] ] in particular @xmath70 . for @xmath71 , let @xmath72 denote the path connected component of @xmath73 . if @xmath74 the set of generators of the complex is @xmath75 . if not , let @xmath76\rightarrow \mathbb{r}$ ] be a hamiltonian function with hamiltonian flow @xmath77 such that @xmath78 is transverse to @xmath7 . denote @xmath79 the space of paths connecting @xmath6 with @xmath7 that are orbits of the hamiltonian flow @xmath77 . fixing @xmath73 and @xmath27 , the @xmath44-floer complex is the complex generated by the subset of orbits of the hamiltonian flow @xmath77 , @xmath80 . @xmath81 _ the differential . _ the differential is defined as usual , by counting elements in the moduli spaces of ( perturbed ) pseudo - holomorphic floer strips with a weight given by an element in the integral group ring @xmath44 , following the same construction that @xcite,@xcite , and a similar version in @xcite . denote by @xmath82 the space of @xmath83-compatible almost complex structures on @xmath68 . we choose a generic , time dependent almost complex structure @xmath84}$ ] ( with @xmath85 for all @xmath86 ) . for any pair @xmath87 , denote by @xmath88 the space consisting of maps @xmath89,m)$ ] such that : 1 . the map @xmath90 satisfies the equation : @xmath91 2 . @xmath92 for @xmath58 , 3 . @xmath93 , 4 . the energy of @xmath94 , is bounded : @xmath95 } \vert \frac{\partial u ( s , t)}{\partial s } \vert dsdt < \infty.\ ] ] the space @xmath96 denotes the quotient by the @xmath97-action on @xmath88 given by reparametrization : @xmath98 acts by @xmath99 . denote by @xmath100 and @xmath101 their @xmath102-dimensional component . for a generic choice @xmath103 the spaces @xmath104 and @xmath105 are manifolds @xcite,@xcite . for @xmath106 the space @xmath107 ) = \ { v \in \hat{\mathcal{m}}^{0}(\gamma_{- } , \gamma_{+};h,\mathbf{j } ) \mid [ v ] = [ u ] \in \pi_{2}(m , l_{0 } \cup \gamma_{- } \cup l_{1 } \cup \gamma_{+})\}$ ] is a compact manifold . to involve the integral group ring @xmath44 in the construction , consider the space obtained from @xmath6 by contracting to a point an embedded path @xmath108 \rightarrow l_{0}$ ] , passing through each point in @xmath109 . denote the resulting space by @xmath110 , and note that it has the same homotopy type as @xmath6 . there is a natural map defined as follows : @xmath111 where @xmath112 is the loop space of @xmath113 . the differential of the @xmath44-floer complex is given by : @xmath114t^{\omega(u)}\gamma_{+}.\ ] ] where @xmath115 is a representative of @xmath116 . note that this is well - defined since for a fixed homotopy class @xmath117 $ ] , the space @xmath107)$ ] is an oriented and compact @xmath118-dimensional manifold so the expression in the right side of equation ( [ differential ] ) makes sense ( see section 4 for definition of sign@xmath119 ) . moreover , the homotopy class of the loop defined by the map @xmath120 is invariant under the @xmath97-action . we do not discuss the way we can assign a grading to the floer complex in this setting , but we will do it in the next section when working with exact lagrangians . 1 . when the pair @xmath61 intersect transversely , we set @xmath121 . in this case the moduli spaces are composed by @xmath122-holomorphic strips connecting intersection points . 2 . a proof of the regularity of the moduli spaces @xmath100 in the relative case ( lagrangian boundary condition ) , can be found in ( * ? ? ? * section 3.2.2 ) . the proof follows the arguments in @xcite , adapted to the monotone lagrangian setting . there is no additional difficulty to see @xmath123 in this setting . the proof follows the same argument than in the monotone setting with coefficients ring @xmath66 . the only difference is that if the minimal maslov number @xmath124 for @xmath58 , then we can have @xmath125 . this can be illustrated by the following example from @xcite . . ] considering @xmath126 ( the path defining the base point ) to be the arc connecting @xmath127 to @xmath128 following the counterclockwise orientation . we have @xmath129t^{\omega(w)}\neq 0 $ ] and @xmath130t^{\omega(v)}\neq 0 $ ] , so @xmath125 . to see @xmath123 we show the following equation : @xmath131[\theta ( u '' ) ] t^{\omega(u ' ) + \omega(u'')}=0.\ ] ] since the pair @xmath132 is assumed to be generic and @xmath133 for @xmath58 , the gromov compactness , the gluing theorem and a choice of compatible orientations ( with gluing ) on the 1-dimensional unparametrized moduli spaces imply the following decomposition of the moduli spaces for a fixed homotopy class @xmath117 $ ] : @xmath134)= \bigcup_{\gamma ' , [ u ' ] + [ u '' ] = [ u ] } \hat{\mathcal{m}}^{0}(\gamma_{- } , \gamma';h,\mathbf{j } , [ u'])\times \hat{\mathcal{m}}^{0}(\gamma ' , \gamma_{+};h,\mathbf{j } , [ u'']).\ ] ] in addition , @xmath135\circ\theta$ ] ( the homotopy class off a the loop given by @xmath120 ) , is constant on a fixed class @xmath117 $ ] and observe that @xmath136[\theta ( u '' ) ] = [ \theta(u'\#u'')]$ ] : for a fixed @xmath137 , let @xmath138 be the base point of @xmath139 . the map @xmath126 defines a path connecting @xmath140 with @xmath141 , for any @xmath142 . denote by @xmath143 this path . then we have that @xmath144[\theta ( u '' ) ] = & [ ( w_{\gamma_{-}}(t))^{-1}u'(s,0)w_{\gamma'}(t)][(w_{\gamma'}(t))^{-1}u''(s,0)w_{\gamma_{+}}(t)]\\ = & [ ( w_{\gamma_{-}}(t))^{-1}u'(s,0)\#u''(s,0)w_{\gamma_{+}}(t)]\\ = & [ \theta(u'\#u'')].\end{aligned}\ ] ] since the signed sum of all boundary components of a 1-dimensional manifold is zero , we have that ( [ d^2=0 ] ) is zero for each @xmath145 and therefore @xmath146 . the homology of this complex is denoted by @xmath147 . the complex @xmath148 is defined by considering the sum over all the components @xmath149 , @xmath150 and the corresponding homology is denoted @xmath151 . the resulting homology is independent on the choice of regular pairs @xmath132 . the proof of this is completely analogous to the monotone case . we will recall the proof of the invariance when changing the hamiltonian @xmath27 , that turns out to be equivalent ( by naturality of floer equation ) to the invariance of the homology under hamiltonian perturbations of a the lagrangian @xmath7 , assuming @xmath152 . we follow the exposition in section 3.2 of @xcite . let @xmath153}$ ] be a hamiltonian isotopy with @xmath154 associated to a hamiltonian function @xmath63 . the isotopy @xmath155 induces a map : @xmath156 \mapsto [ \varphi_{t}(\gamma(t))].\ ] ] assume that @xmath152 and @xmath157 . fix the data @xmath158 . to compare the homology of the two complexes @xmath159 and @xmath160 , a chain map @xmath161 is defined using moving boundary conditions . the moduli spaces with moving boundary conditions were introduced by oh in @xcite . in order to define this map we need to introduce some notation . let @xmath162 $ ] , be a smooth function such that @xmath163 for @xmath164 , @xmath165 for @xmath166 and @xmath167 is strictly increasing in @xmath168 . for @xmath169 and @xmath170 , consider @xmath171 to be the moduli space of maps @xmath89,m)$ ] such that : 1 . for all @xmath24 , @xmath172 and @xmath173 , 2 . @xmath174 , 3 . 4 . @xmath176 . let @xmath177 and fix @xmath141 a path in the connected component @xmath73 . define the functional : @xmath178 for each map @xmath179 let @xmath180\longrightarrow m$ ] denote the map defined by @xmath181 . the chain map is defined as follows : @xmath182t^{\omega(v_{u})- \phi_{g}(\gamma_{+ } ) } \big)\gamma_{+}.\ ] ] in order to obtain a loop @xmath183 , we consider the space @xmath184 as in the previous section , where the embedded path @xmath108 \rightarrow l_{0}$ ] passes through each point in @xmath185 . to see that the map @xmath186 is a chain map we compare the following equations : @xmath187[\theta(w ) ] t^{\omega(v_{u})+\omega(w)-\phi_{g}(\gamma')}\\ & \langle\widetilde{c}_{\varphi}(d(\gamma_{-})),\gamma_{+}\rangle = \\ & \sum_{\substack { \gamma '' \in l_{0}\cap l_{1}\\(\hat{u'},w')\in \hat{\mathcal{m}}^{0}(\gamma_{- } , \gamma'')\times \mathcal{m}_{\varphi}^{0}(\gamma '' , \gamma_{+ } ) } } \text{sign}(u')\text{sign}(w')[\theta(u')][\theta(w ' ) ] t^{\omega(v_{w'})+\omega(u')-\phi_{g}(\gamma_{+})}.\end{aligned}\ ] ] for generic @xmath188 and fixed homotopy class @xmath189 , the space @xmath190 admits a compactification into a @xmath191-dimensional manifold @xmath192)$ ] , with boundary given by : @xmath193 + @xmath194 + @xmath195 the choices of orientations and spin structures on @xmath4 induce canonical orientations on this spaces , compatible with the gluing map . moreover , any two boundary components @xmath196 of @xmath197)$ ] in the same connected component satisfy @xmath198[\theta(w ) ] = [ \theta(u')][\theta(w')]$ ] and @xmath199 therefor , the sum of the two expressions above vanishes . in a similar way we define a chain map @xmath200 such that @xmath201 is a quasi - inverse of @xmath202 , concluding that the map ; @xmath203 is an isomorphism . we summarize this section in the following proposition : [ prophf ] let @xmath61 be a pair of oriented and spin monotone lagrangians , equipped with a choice of spin structure . assume that the minimal maslov number satisfies @xmath204 for @xmath58 . let @xmath103 be generic , then the floer complex @xmath205 is a chain complex . + moreover , the homology @xmath206 does not depend on the choice of generic pair @xmath103 and is invariant under hamiltonian isotopies of @xmath6 or @xmath7 . proposition [ prophf ] is a @xmath44-version of the same result due to oh @xcite in the monotone setting . for twisted coefficients , the @xmath207-version of proposition [ prophf ] appears in @xcite for a pair @xmath61 of non - compact lagrangians with @xmath208 and @xmath209 . in the monotone setting , a similar result for the lifted floer homology appears in @xcite . in this subsection we adapt the chain complex defined in the last section in order to associate such a complex to a lagrangian cobordism @xmath0 . instead of working with a lagrangian cobordism we use the corresponding non - compact lagrangian with cylindrical ends ( obtained when extending the ends of the cobordism ) . the floer theory for lagrangian with cylindrical ends was developed in @xcite . given a symplectic manifold @xmath2 , denote by @xmath210 the symplectic manifold composed by @xmath211 and @xmath212 . denote by @xmath213 the projection @xmath214 . * section 4.1 ) an ( elementary ) lagrangian submanifold with cylindrical ends , @xmath215 is a lagrangian submanifold without boundary and with the following properties : 1 . for every @xmath216 the subset @xmath217\times \mathbb{r})\times m$ ] is compact . 2 . there exists @xmath218 with @xmath219 such that : @xmath220 @xmath221\times \mathbb{r})\times m = ( ( -\infty , r_{-}]\times \{a_{-}\})\times l_{0}\ ] ] for some pair of lagrangian @xmath222 and @xmath223 . we use the term elementary since the general definition appearing in @xcite , considers lagrangians with more than one positive or negative end . for every @xmath224 , let @xmath225 denote the positive cylindrical end of @xmath33 , and similarly for @xmath226 , @xmath227 will denote the negative cylindrical end of @xmath228 let @xmath229 be two lagrangians with cylindrical ends . we say that they are cylindrically distinct at infinity if there exists @xmath230 such that @xmath231 and @xmath232 any lagrangian cobordism @xmath0 extends to a lagrangian with cylindrical ends @xmath33 in the following way;@xmath233\times \{1\}\times l_{0 } ) \cup w \cup ( [ 1 , \infty)\times \{1\}\times l_{1}).\ ] ] in the class of lagrangians with cylindrical ends we consider hamiltonian isotopies that `` fix '' the cylindrical ends in the sense of the following definition : ( * ? ? ? * definition 4.1.2 ) an isotopy @xmath234}$ ] of lagrangian submanifolds with cylindrical ends of @xmath235 is called horizontal isotopy if there exists a hamiltonian isotopy @xmath236}$ ] of @xmath235 , with @xmath237 and the following properties : * @xmath238 for all @xmath239 $ ] . * there exist real numbers @xmath240 and a constant @xmath241 , such that for all @xmath239 $ ] and @xmath242 , we have @xmath243 . * for all @xmath242 , @xmath244 here , @xmath245 is the time dependent vector field of the flow @xmath236}$ ] . we now proceed to define the @xmath246-floer complex associated to a pair of monotone lagrangians with cylindrical ends @xmath247 . consider @xmath247 a pair of monotone lagrangians with cylindrical ends with @xmath248 for @xmath58 . in @xcite it is shown that the floer homology with @xmath249-coefficients , @xmath250)$ ] is well defined and depends on an additional data @xmath251 $ ] , coming from the choice of a hamiltonian perturbation of the ends of @xmath252 , which makes the image of @xmath252 under this perturbation cylindrically distinct from @xmath253 at infinity . when working with @xmath254-coefficients , the construction is completely analogous under additional choices of spin structures on the pair @xmath247 . the construction of the floer complex follows the schema presented in the previous section . given the non - compactness of the lagrangian pair , additional choices have to be made to ensure the compactness of the moduli spaces . the floer complex is defined using the following data : * a fixed component @xmath255 . * a perturbation @xmath256 where @xmath257\times\widetilde{m}\rightarrow \mathbb{r}$ ] and @xmath258 are two hamiltonians satisfying : 1 . @xmath27 has compact support . 2 . for @xmath259 big enough so that @xmath252 is cylindrical on @xmath260 , the function @xmath261 is such that the support of @xmath261 is contained in a neighborhoods @xmath262 of @xmath263 where @xmath264 with @xmath265 . + we denote the space of pairs @xmath256 as above by @xmath266 . * an almost complex structure @xmath267 on @xmath210 . we will restrict to a family of time dependent complex structure @xmath268 where @xmath269 is a compact set , and @xmath270}\in \mathcal{\widetilde{j}}_{b}$ ] satisfies the following properties : 1 . for every @xmath86 , @xmath271 is an @xmath272-tamed almost complex structure on @xmath235 . 2 . for every @xmath86 , the projection @xmath213 is @xmath273-holomorphic on @xmath274 . if @xmath275 , we write @xmath276 . the additional choice of perturbation @xmath277 guaranties that @xmath253 and @xmath278 are cylindrically distinct at infinity . if the ends of @xmath253 coincide with those of @xmath252 , the space @xmath266 has four connected components fixed by the four possible choices of the function @xmath261 ( depending on the signs of @xmath279 ) . we denote by @xmath251 $ ] the path component of @xmath266 associated to a perturbation @xmath256 . the choice of almost complex structure @xmath280 implies the compactness of the moduli spaces of @xmath122-holomorphic strips ( in a fixed relative to the boundary homotopy class ) , since an application of the open mapping theorem shows that any pseudo - holomorphic curve with finite energy has its image in a fixed compact set . for the space of perturbed @xmath122-holomorphic strips , with boundary on a lagrangian pair @xmath281 cylindrically distinct at infinity , we apply a hamiltonian perturbation @xmath282 to the second lagrangian @xmath252 , such that the pair @xmath283 intersect transversely . after a naturality argument we can conclude that the moduli spaces of perturbed @xmath122-holomorphic strips is compact . for a detailed proof of compactness in the cobordism setting , see ( * ? ? ? * lemma 4.2.1 ) . for generic data @xmath284 , the floer complex @xmath285 is well - defined ( * ? ? ? * section 4.3 ) . as in previous section , @xmath286 denotes the sum along all the class of hamiltonian chords @xmath287 . from now on , we make no distinction in the notation of a cobordism @xmath0 and its associated lagrangian with cylindrical ends @xmath33 , unless it is necessary . [ prop ] let @xmath288 be a pair of oriented and spin monotone lagrangians with cylindrical ends equipped with a choice of spin structure . assume that the minimal maslov number satisfies @xmath289 for @xmath58 . let @xmath284 be generic . then the floer complex @xmath290 is a chain complex . moreover , its homology @xmath291 does not depend on the choice of generic pair @xmath292 but depends on the path connected component @xmath251 \in \pi_{0}(\mathcal{h}(w_{0 } , w_{1}))$ ] and is invariant under horizontal isotopies of @xmath293 or @xmath294 , up to an isomorphism . once the compactness of the moduli spaces is guarantied , the proof of this proposition follows from last section . the only additional argument concerns the invariance of the homology for different perturbations @xmath295 with @xmath251 = [ f']$ ] . this issue is addressed in the proof of the analogous statement for @xmath249-coefficients , ( * ? ? ? * proposition 4.3.1 ) . the idea is to study a chain map defined using moving boundary conditions induced by an homotopy @xmath296 , connecting @xmath261 , with @xmath297 . the homology of the complex @xmath290 is denoted by @xmath298).\ ] ] this theorem is a consequence of proposition [ prophf ] and proposition [ prop ] . let @xmath299 be an orientable , spin , monotone , lagrangian cobordism viewed as a lagrangian with cylindrical ends . let @xmath300 be an element in the set @xmath301 . consider @xmath302 a perturbation where @xmath36 is locally given by @xmath303 with the property that @xmath304 if @xmath305 and @xmath306 if not , as well as @xmath307 if @xmath308 and @xmath309 if not . such a perturbation makes @xmath5 and @xmath310 cylindrically distinct at infinity . denote by @xmath311 the lagrangian @xmath312 . consider the floer complex @xmath313 . from proposition [ prop ] follows that + @xmath314)$ ] depends only on @xmath35 $ ] . for @xmath315 , notice that there is an horizontal isotopy @xmath316 such that @xmath317 : ] in a similar way , when @xmath318 we can also find a horizontal isotopy @xmath319 such that @xmath320 . thus , @xmath321 from the invariance of the homology under horizontal isotopies ( proposition 2.2 ) , we can conclude that @xmath322 ) \cong 0 \cong hf(w , w;\mathbb{z}(\pi_{1}w),[f_{l_{1}}]).\ ] ] consider now the case when @xmath323 , ( the case @xmath324 is completely analogous ) . notice that there are horizontal isotopies in the same path component @xmath325 ( see figure ) , such that : @xmath326 and in the same way @xmath327 ] from the invariance of the homology under horizontal isotopies ( proposition 2.2 ) , we can conclude that @xmath328 moreover , the relative homology is independent of the choice of almost complex structure . denoting by @xmath329 the homology of the complex @xmath330 , for @xmath20 , we have that @xmath331 in this section we apply the previous results to the class of exact lagrangian cobordisms . let @xmath2 be a symplectic manifold where the two - form @xmath332 is exact . if the one - form restricted to a lagrangian @xmath48 is also exact @xmath333 , then we say that the lagrangian @xmath48 is an exact lagrangian . an exact lagrangian cobordism is a lagrangian cobordism @xmath0 where @xmath57 is an exact lagrangian . for an exact lagrangian submanifold @xmath28 the floer homolgy @xmath334 is known to be isomorphic to the morse homology @xmath335 . here @xmath336 denotes the morse complex of @xmath48 , @xmath337 denotes a morse - smale pair composed by a morse - smale function @xmath261 and a riemannian metric @xmath338 . we can expect the @xmath339-floer homology @xmath340 to be related to the homology of the @xmath339-coefficients morse complex @xmath341 . * definition 1.4)the @xmath339-coefficients morse complex @xmath342 is the based finite generated @xmath339-module chain complex given by @xmath343 here @xmath344 denotes the pullback of @xmath345 and @xmath346 denotes the signed sum of the number of connecting flow lines of @xmath344 between @xmath347 and @xmath348 . notice that the homology of the @xmath339-coefficients morse complex @xmath349 is isomorphic to the singular homology of the universal covering of @xmath48 , denoted by @xmath350 . this follows from the observation that the lifted morse - smale function defines a cellular decomposition on @xmath351 . let @xmath352 be an orientable and spin exact lagrangian submanifold . then @xmath353 . before we start the proof of this proposition , we will define the version of the floer complex for exact lagrangians used here . for this section we follow the exposition in @xcite . _ coefficient ring . _ in the definition of the floer complex we have used the universal novikov ring @xmath66 , when working with exact lagrangians this ring is not necessary anymore . consider a pair of exact lagrangians @xmath354 . there are functions @xmath355 such that @xmath356 for @xmath58 . the symplectic area @xmath357 of any strip with boundary conditions on @xmath358 connecting two intersection points @xmath359 , is given by @xmath360 . then , there are uniform bounds on the energy in a fixed moduli space @xmath361 , independently of the homotopy class of the strip @xmath117 \in \pi_{2}(m , l_{0 } \cup l_{1}\cup x \cup y)$ ] . thus , is not necessary to consider the universal novikov ring @xmath66 in this case . the module used in the definition of the floer complex @xmath362 is + @xmath363 . here , @xmath364 is the image of @xmath6 under the time - one hamiltonian diffeomorphism associated to @xmath27 , such that @xmath74 . _ grading . _ in the module @xmath363 , there is a natural graduation defined using the maslov - viterbo index . let @xmath365 \rightarrow m$ ] be a strip with boundary conditions on @xmath4 joining two intersection points @xmath366 . the maslov - viterbo index of @xmath90 , denoted by @xmath367 , is the maslov index of a loop obtained from a trivialization of the bundle @xmath368 . let @xmath369 \rightarrow l_{0}$ ] , @xmath370 \rightarrow l_{1}$ ] be two paths defined by @xmath371 and @xmath372 . consider the path of lagrangians @xmath373 concatenated with a path of lagrangians @xmath374 which intersect transversely @xmath375 at each @xmath24 . the maslov - viterbo index of @xmath90 , is the maslov index of the loop @xmath376 after trivialization . fix a point @xmath377 . consider the path - connected component of @xmath378 in @xmath379 denoted by @xmath380 . for any other point @xmath381 we define @xmath382 to be the maslov - viterbo index of a path @xmath383 in @xmath380 connecting @xmath128 with @xmath378 . this is independent on the choice of path @xmath383 since the condition @xmath384 implies that , for a small hamiltonian perturbation , all floer strips defining the floer homology are contained in a small tubular neighborhood of @xmath6 . thus they can be seen as strips on the cotagent bundle @xmath385 . for the cotangent bundle the maslov index is independent on the homotopy class of the strip . assuming that @xmath386 , we set @xmath387 . by the weinstein theorem there exists a neighborhood of @xmath48 , @xmath388 symplectomorphic to a tubular neighborhood of the zero - section . let @xmath389 be a morse function and for a small @xmath9 let @xmath390 . note that @xmath391 is hamiltonian isotopic to @xmath48 . the set of intersection points @xmath392 is the set @xmath393 . the relation between the floer and morse grading is given by @xmath394 , where @xmath395 denotes the morse index ( see@xcite ) . this implies that the grading of the floer and the morse complexes agree up to a shift . moreover , from @xcite there exist a time dependent almost complex structure @xmath188 for which the moduli spaces of floer strips correspond to moduli spaces of connecting flow lines . on the other hand , in ( * ? ? ? * theorem 13 , appendix b ) was shown that the orientations of the morse complex obtained from the geometrical constructions can be extended to canonical orientations induced by the determinant bundle , the last ones are the ones used to define the floer complex . then @xmath396 . here , @xmath397 denotes the @xmath339-coefficients morse complex . to see the last equality @xmath398 , notice that for @xmath399 , @xmath400y = & \sum _ { y \in crit(f)_{*-1 } } ( \sum _ { [ \theta(u ) ] \in\rho(\theta(\mathcal{m}^{0}(x , y ) ) ) } n^{\tilde{f } , \tilde{g}_{\mathbf{j}}}(\tilde{x } , [ \theta(u)]\tilde{y } ) [ \theta(u)]\big ) y \\ = & \sum _ { y \in crit(f)_{*-1 } } ( \sum _ { \alpha \in \pi_{1}l } n^{\tilde{f } , \tilde{g}_{\mathbf{j}}}(\tilde{x } , \alpha\tilde{y } ) \alpha ) y.\end{aligned}\ ] ] here @xmath401 denotes the projection . now , from the invariance under hamiltonian perturbations , and the independence on the choices defining the floer homology , we can conclude that @xmath402 we now consider the exact lagrangian with cylindrical ends associated to an exact lagrangian cobordism @xmath14 . for such a lagrangian the floer homology depends on the choice of path component @xmath251 $ ] , defining the perturbation on the cylindrical ends . different choices lead to different homologies . in order to relate the floer homology @xmath403)$ ] with the singular homology , let @xmath300 be as in theorem [ teorema ] , an element of in the set @xmath301 . denote by @xmath35 $ ] the class of perturbation with @xmath304 if @xmath305 and @xmath306 if not . as well as @xmath307 if @xmath308 and @xmath309 if not . let @xmath404 be an autonomous almost complex structure that splits @xmath405 ( for some @xmath406 ) and denote by @xmath407 its induced riemannian metric . let @xmath408 be a morse function such that the negative gradient of @xmath409 is transverse to @xmath410 and it points outside of @xmath5 along @xmath300 and inside of @xmath5 on the complement of @xmath300 in @xmath410 . [ propcob ] let @xmath14 be an orientable and spin exact lagrangian cobordism equipped with a choice of spin structure . we have the following isomorphisms:@xmath411 ) \cong h(cm(w;\mathbb{z}(\pi_{1}w),(\widetilde{f}_{s } , g_{\tilde{j}}))).\ ] ] extend the morse function @xmath412 to a function @xmath413 such that there is a compact set @xmath414 where @xmath415 is a linear function . consider a weinstein neighborhood @xmath416 of @xmath33 , where we see @xmath33 as the zero - section in @xmath417 with the standard exact symplectic form . let @xmath418 be the image of @xmath33 under the time one hamiltonian diffeomorphism in the hamiltonian isotopy @xmath419 , generated by the hamiltonian @xmath420 . here , @xmath421 . the pair @xmath422is cylindrically distinct at infinity and intersects transversely , since the hamiltonian flow @xmath423 is a translation outside some compact region . from @xcite , there exist a time dependent almost complex structure @xmath267 , such that there is a bijection between the moduli space of floer strips and the moduli space of flow lines connecting two critical points @xmath424 . condition 2 on the morse function @xmath425 ensures that @xmath426 . since the hamiltonian isotopy is linear outside the set @xmath274 then @xmath427 on this set . from the identity @xmath428 follows the claim . then , the floer complex @xmath429 is well defined and @xmath430 notice that if @xmath431 or @xmath358 , then the connected path @xmath35 $ ] corresponds to perturbations @xmath264 with @xmath432 and @xmath433 of opposite signs . using the invariance of the @xmath13-floer homology under horizontal hamiltonian perturbations ( theorem [ teorema ] and proposition 2.2 ) , we have the following isomorphisms : @xmath434)\cong hm_{*}(l_{1};\mathbb{z}(\pi_{1}w))\cong h_{*}(\tilde{l}_{1}).\ ] ] in the case @xmath435 , the invariance of the @xmath13-floer homology under horizontal hamiltonian perturbations ( theorem [ teorema ] and the choice of morse function on the cobordism imply : @xmath436)\cong h_{*}(cm(w;\mathbb{z}(\pi_{1}w),(\widetilde{f}_{l_{i } } , g_{\tilde{j}})))\cong h_{*}(\tilde{w } , \tilde{l}_{i } ) = 0.\ ] ] for the relation between singular and @xmath13-coefficients morse homology see @xcite . [ col1 ] if @xmath14 is an orientable and spin , exact lagrangian cobordism and @xmath437 is an isomorphism for @xmath41 , then @xmath0 is an h - cobordism . [ col2 ] if @xmath0 is an orientable and spin , exact lagrangian cobordism with @xmath42 and @xmath43 ( @xmath41 ) , @xmath5 are simply connected , then @xmath0 is a lagrangian pseudo - isotopy . to prove these corollaries we first recall the @xmath438-cobordism theorem . if @xmath0 is a simply connected h - cobordism with dim@xmath439 , then the cobordism @xmath0 is diffeomorphic to the trivial cobordism @xmath22 \times l_{0 } ; l_{0 } , l_{0})$ ] . from the remark below , we have that if @xmath0 is an exact lagrangian cobordism , orientable and spin , equipped with a choice of spin structure then @xmath440 for @xmath41 . here , the homology @xmath441 is the relative homology induced from a cw - decomposition of the pair @xmath442 . the hypothesis on the fundamental groups @xmath437 combined with @xmath440 for @xmath41 , imply that the inclusions are homotopy equivalences . this follows from one of the whitehead theorems @xcite theorem 3 . from this we deduce corollary [ col1 ] . for the second corollary we use corollary [ col1 ] and the h - cobordism theorem . this section presents the results on simple homotopy theory used in the paper . we follow the exposition of cohen @xcite . the simple homotopy theory is the study of the relation of simple homotopy type between cellular complexes . let @xmath443 and @xmath48 be finite cw - complexes . there is an elementary collapse from @xmath443 to @xmath48 ( or an elementary expansion from @xmath48 to @xmath443 ) if @xmath444 , where : 1 . @xmath445 is the attaching map for @xmath446 and @xmath447 is the attaching map for @xmath448 . 2 . the new cells attached by @xmath261 are not in @xmath48 . 3 . the closure of @xmath449 denoted by @xmath450 , satisfies @xmath451 . two cw - complexes @xmath48 and @xmath443 have the same _ simple - homotopy type _ if they are related by a finite sequence of collapses and expansions , denoted @xmath452 . a map @xmath453 is a _ simple - homotopy equivalence _ if @xmath261 is homotopic to a map obtained by a finite number of compositions of the maps induced by collapses and expansions . the relation of simple homotopy type is finer than that of homotopy type . there are cw - complexes that are homotopy equivalent but not simple homotopy equivalent , examples can be found in @xcite . given a fixed cw - complex @xmath48 , consider the set of pairs @xmath454 , where @xmath443 is a cw - complex homotopy equivalent to @xmath48 , that contains @xmath48 as a sub - complex . the whitehead group of @xmath48 , @xmath455 , is defined to be the set of equivalence classes of pairs @xmath456 $ ] under the relation @xmath457 if @xmath458 rel @xmath48 . this means that @xmath443 and @xmath459 are related by a finite set of collapses and expansions for which no cell of @xmath48 is ever removed . the group operation is @xmath460 + [ k ' , l ] = [ k\cup_{l } k ' , l].\ ] ] where @xmath461 is the disjoint union of @xmath443 and @xmath459 identified by the identity map on @xmath48 . the whitehead group of @xmath48 , @xmath455 , is an abelian group ( see 6.1 of @xcite ) . given a homotopy equivalence @xmath462 , we define its whitehead torsion @xmath463 $ ] , where @xmath464 denotes the mapping cylinder of @xmath261 . the whitehead torsion is the obstruction for a homotopy equivalence to be a simple - homotopy equivalence as stated in the following theorem . a homotopy equivalence @xmath462 , is a simple homotopy equivalence if and only if @xmath465 . the geometric definitions of whitehead group is hard to work with , instead there is an algebraic version that we introduce in the following subsection . let @xmath62 be the integral group ring of the group @xmath63 . the group @xmath466 denotes the group of non - singular @xmath467 matrices over @xmath62 . identifying each @xmath468 with the matrix @xmath469,\ ] ] we obtain an injection @xmath470 . the infinite general linear group of @xmath62 is the union over all @xmath102 of the groups @xmath471 and is denoted by @xmath472 . a matrix in @xmath472 is called elementary if it coincides with the identity matrix except for one off - diagonal element . the set of elementary matrices is the commutator subgroup of @xmath472 . let @xmath473 denote the group generated by all the elementary matrices and all matrices that coincide with the identity matrix except for one diagonal element @xmath474 . the whitehead group of @xmath63 is the group : @xmath475 the class of a matrix @xmath476 \in wh(g)$ ] is called the whitehead torsion of the matrix and is denoted by @xmath477 . a @xmath62-module is a free @xmath62-module @xmath68 with a distinguished family of bases @xmath478 which satisfy : if @xmath479 and @xmath480 are two bases of @xmath68 and if @xmath481 , then @xmath482 if and only if @xmath483 ) = 0 \in wh(g)$ ] . here @xmath484 $ ] represents the matrix that changes the base @xmath479 to the base @xmath480 . an isomorphism @xmath485 is a _ simple isomorphism of @xmath62-modules _ if @xmath486 , where @xmath487 is the matrix of @xmath261 with respect to any distinguished basis of @xmath488 . a @xmath62-complex is a free chain complex over @xmath62 , @xmath489 such that each @xmath490 is a @xmath62-module . a preferred basis of @xmath491 mean a basis @xmath492 where @xmath493 is a preferred basis of @xmath494 . a _ simple isomorphism _ of @xmath62-complexes , @xmath495 , is a chain mapping such that @xmath496 is a simple isomorphisms for all @xmath497 . another useful notion is the following . two @xmath62-complexes , @xmath498 and @xmath499 , are _ simple homotopy equivalent _ , @xmath500 , if there exist trivial - complex @xmath501 is elementary trivial if @xmath502 . a trivial chain complex is a direct sum of elementary trivial chain complexes . ] @xmath62-chain complexes @xmath503 and @xmath504 such that the chain complex @xmath505 is simple isomorphic to the chain complex @xmath506 . let @xmath498 be an acyclic @xmath62-complex . then @xmath498 is contractible , so there exist a chain contraction @xmath507 $ ] such that @xmath508 . if @xmath509 and @xmath510 , then the torsion of the complex @xmath498 is defined to be the torsion of the map : @xmath511 @xmath512 the whitehead torsion of a complex satisfies the following properties : 1 . if @xmath513 is a simple isomorphism of chain @xmath62-complexes , then @xmath514 . 2 . the torsion of a direct sum of @xmath62-complexes @xmath515 satisfies @xmath516 . if @xmath513 is a homotopy equivalence of @xmath62-complexes , then the whitehead torsion of @xmath261 is defined by @xmath517 . let @xmath518 be a pair of finite , connected cw - complex where @xmath48 is a sub - complex of @xmath5 . let @xmath519 denote the cellular complex , where @xmath520 is the free module generated by the @xmath497-cells of @xmath521 and the differential is defined by the boundary operator on the exact sequence for singular homology of the triple @xmath522 . if @xmath523 is a homotopy equivalence , then the induced map @xmath524 is an isomorphism . in this case , if @xmath525 is the universal covering of @xmath5 , then @xmath526 is the universal covering of @xmath48 and @xmath527 is also a homotopy equivalence . the projection map @xmath528 induces a structure of cw - complex on @xmath529 from the ones on @xmath530 . with this cellular structure , the cellular complex @xmath531 is a @xmath339-complex . let @xmath518 be a pair of finite , connected cw - complexes , and suppose that @xmath523 is a homotopy equivalence . the whitehead torsion of the pair is defined by @xmath532 where @xmath533 is the universal covering of @xmath534 . one application of the theory , briefly recalled in this section , is the s - cobordism theorem . let @xmath0 be an h - cobordism of dimension @xmath535 . then the torsion @xmath536 vanishes if and only if @xmath5 is diffeomorphic to the product @xmath537 $ ] . notice that if @xmath0 is an h - cobordism , then we can also define the whitehead torsion of a pair @xmath538 using a morse function . recall that any nice morse function defined on an h - cobordism @xmath0 induces a cw - structure on it , that lifts to the universal covering @xmath539 . the following theorem of milnor establishes that the torsion of the complex defined by the cw - pair induced by the lift of any nice morse function is the torsion of the pair @xmath540 . if @xmath541 is an h - cobordism , then @xmath542 does not depend on the choice of nice morse function . [ remarckwh ] recall that under some conditions established in corollary [ col1 ] , if @xmath0 is an exact lagrangian cobordism , then it is an h - cobordism and the torsion of the pair @xmath540 is well defined . from milnors s theorem follows that @xmath543 , for a pair @xmath544 satisfying the conditions required in the proof of proposition 2.4 . by the s - cobordism theorem , an h - cobordism of dimension higher than 5 is a pseudo - isotopy if its whitehead torsion vanishes . in the symplectic category , the cobordism studied has an additional structure : it is a lagrangian submanifold with lagrangian boundary . to compute the whitehead torsion of an exact lagrangian h - cobordism @xmath0 , we study the torsion of the @xmath13-floer complex associated to the pair of exact lagrangians with cylindrical ends @xmath545 , where @xmath546 satisfies @xmath547 . in this section we show that the whitehead torsion of the @xmath13-floer complex is invariant under horizontal perturbations of the lagrangian with cylindrical ends associated to the cobordism . fukaya defined a whitehead torsion for the floer complex of lagrangian intersections and he suggested the invariance of it under hamiltonian perturbations in @xcite . he proposed the following conjecture . if @xmath548 and @xmath549 , then there exists an exact symplectic diffeomorphism @xmath550 such that @xmath551 . m. sullivan @xcite proved the invariance of the whitehead torsion of the @xmath552-coefficients floer complex under compactly supported hamiltonian isotopies . he considered the @xmath552-coefficients floer complex associated to a pair of hamiltonian isotopic , non - compact lagrangians @xmath553 , such that @xmath554 . part of his proof is based on a gluing theorem proved by lee . lee studied the floer - novikov complex for periodic orbits of a hamiltonian . in @xcite and @xcite , she defined an invariant , denoted by @xmath555 , related to the reidemeister torsion of the floer - novikov complex . based on a series of gluing theorems , she proved that @xmath555 is invariant under hamiltonian perturbation . her work is a generalization to the symplectic setting of the same invariant @xmath555 , studied by hutchings and her for the novikov complex in @xcite . from the theory exposed in previous section , any finite acyclic @xmath13-complex with a preferred family of bases has a well defined whitehead torsion . the @xmath13-floer complex associated to the pair @xmath545 has a preferred base given by the set @xmath556 , and is an acyclic complex for the class of perturbations that displace @xmath311 away from @xmath5 ; therefore it has a well - defined whitehead torsion . from remark ( [ remarckwh ] ) , the whitehead torsion of an exact lagrangian h - cobordism , can be defined using the @xmath13-morse complex . the @xmath13-morse complex agrees with the @xmath13-floer complex for a special choice of hamiltonian . the invariance of the whitehead torsion of the @xmath13-floer complex under horizontal perturbations , implies that the whitehead torsion of an exact lagrangian h - cobordism is given by the whitehead torsion of its @xmath13-floer complex . as an application of this , we establish that an exact lagrangian cobordism ( satisfying appropriate hypotheses ) , is a lagrangian pseudo - isotopy . the main results of the section are the following : [ theorem invariance ] let @xmath545 be a pair of exact lagrangians with cylindrical ends where @xmath557 for @xmath435 and @xmath547 . consider the @xmath13-floer complex @xmath558 , where @xmath559 . if @xmath560}$ ] is a horizontal isotopy , then the whitehead torsion of the floer complex , satisfies : @xmath561 and the main result of the paper : [ s - theorem ] let @xmath0 be an exact , orientable and spin lagrangian cobordism equipped with a choice of spin structure . assume @xmath29 . if the map @xmath30 induced by the inclusion @xmath31 is an isomorphism for @xmath32 , then @xmath0 is a lagrangian pseudo - isotopy . the technique of bifurcation analysis is used for proving the invariance of floer / morse complexes under some parameter . it is also used to prove the independence under hamiltonian perturbation of torsion - type invariants associated to floer / morse complexes . in this subsection we explain the bifurcation analysis technique . the idea of this technique is to consider a generic homotopy joining the two generic parameters defining two floer complexes . at each time of the homotopy the parameters associated have some regularity . at certain times along the homotopy regularity the technique consists in describing how the complex changes after each one of these moments called bifurcations . floer s first proof of the independence of the floer homology under hamiltonian perturbations , used bifurcation analysis @xcite . through a series of gluing theorems , he describes the changes of the floer complex after a bifurcation occurs . the complete proofs of these gluing theorems , for the floer complex , appeared in lee @xcite , @xcite . basic bifurcation theory appears in the morse theory . given two morse - smale pairs @xmath562 and @xmath563 ( @xmath564 is a morse function and @xmath565 a morse - smale pseudo - gradient vector field for @xmath566 ) , consider a homotopy joining these parameters @xmath567 . from cerf s work @xcite , it is known that in a generic homotopy , for a finite number of parameters , the pair @xmath568 is not morse - smale . the two phenomena that generically occur at isolated times are : 1 . there is a degenerated critical point , that has only one degenerate direction with a quadratic tangency , where two non - degenerated critical points are born or die . this phenomenon is called birth - death . 2 . there are two non - degenerated critical points of same index for which the unstable and stable manifolds do not intersect transversely , yielding a degenerate flow line . this phenomenon is called handle - slide . for the floer complex of lagrangian intersections , when the lagrangians are exact , the bifurcations occurring along a homotopy joining two generic set of parameters are analog to the ones occurring to the morse complex . fix a pair of exact lagrangians with cylindrical ends @xmath569 . the data @xmath570 , where @xmath571 and @xmath188 is a time - dependent , compatible almost complex structure is called generic if the associated floer complex + @xmath572 is well defined , namely : 1 . the lagrangians intersect transversely , @xmath573 . 2 . for every @xmath188-holomorphic strip @xmath90 , the linearization of the time dependent cauchy - riemann operator @xmath574 is surjective . let @xmath575 $ ] and consider a homotopy @xmath576 joining two generic data @xmath577 and @xmath578 . in general , the floer complex is not well defined at all times of the homotopy . a bifurcation along the homotopy @xmath579 is a time @xmath580 such that @xmath581 is not generic . to study the changes of the whitehead torsion of the @xmath13-floer complex , we will pick a nice homotopy which will allow as to compare the two complexes and show that they are simply homotopy equivalent . the homotopies that we will consider are defined as follows : a generic homotopy @xmath582 joining generic data @xmath583 and @xmath584 , is a smooth homotopy such that the only bifurcations occur at isolated times and are of the following type : 1 . births - deaths : at some time @xmath585 in the homotopy there is a non - transverse intersection point in @xmath586 , such that it has only one degenerate direction with a quadratic tangency , where two non - degenerated intersection points appear or disappear . 2 . handle slides : a degenerate strip @xmath90 of index @xmath587 between non - degenerate intersection points . the following proposition assure us that generic homotopies exist . let @xmath588 denote the set of horizontal isotopies @xmath589 joining @xmath590 with @xmath423 and denote by @xmath591 the space of one - parameter families of time dependent almost complex structures @xmath592 joining @xmath593 with @xmath594 , where @xmath595 for each @xmath596 . given two generic data @xmath583 and @xmath584 , there is a non - empty set @xmath597 of regular homotopies @xmath582 joining them . the proof appears in ( * ? ? ? * lemma 3.3 ) and ( * ? ? ? * proposition 3.2 ) . floer s proof still applies to this non - compact setting , since the main argument is based on the structure and regularity of the spaces of functions @xmath588 and of almost - complex structures @xmath591 , and the regularity of these spaces is independent of the compactness of the setting . consider a regular homotopy @xmath582 joining generic data @xmath583 and @xmath584 . let @xmath598 be the set of parameters where a bifurcation occurs . a regular homotopy defines a one - parameter family of @xmath13-floer complexes , @xmath599 . the proof of theorem [ theorem invariance ] consists in the observation that through a regular homotopy , the bifurcations that the @xmath13-floer complexes suffer are elementary ( change of basis or sum with a trivial complex ) . then , the @xmath13-floer complex associated to @xmath583 is simply homotopy equivalent to the @xmath13-floer complex associated to @xmath584 . this implies that the whitehead torsion of the @xmath13-floer complex is invariant under hamiltonian perturbations . in the following subsections we describe the behavior of the @xmath13-floer complex after the two types of bifurcations , birth - deaths and handle - slides . to do this , let us assume that @xmath600 is a regular homotopy joining two generic sets of data @xmath601 with @xmath602 . we will assume that exactly one of the two phenomena at time @xmath603 occurs , and we will describe the changes in @xmath604 with respect to @xmath605 . in this subsection we study the birth - type bifurcation following @xcite ; the analysis for a death type bifurcation is completely analogous . assume that there is a birth at @xmath585 . thus , there is a non - transverse intersection point @xmath606 , which gives birth to two families of non - degenerated intersection points denoted by @xmath607 , and @xmath608 with @xmath609 where @xmath610 . in this situation the @xmath13-floer complex changes in the following way : 1 . there are two new generators for the module @xmath611 , namely @xmath612 there is a new moduli space @xmath613 , composed by strips connecting the two new generators . 2 . new strips connecting old intersection points to the new ones can appear . the spaces @xmath614 and @xmath615 where @xmath616 and @xmath617 could be non - empty . this will affect the differential of the @xmath618 complex . we say that a birth or a death is independent , if the last two spaces are empty . an independent birth / death does not change the simple homotopy type of the complex , and thus the whitehead torsion . see also lemma 4.5 . a stabilization technique of m. sullivan @xcite , reduces the general case of a birth / death to that of independent birth / deaths as we explain below . the crucial point here is the existence of a regular homotopy connecting the data @xmath619 with @xmath620 , such that after applying the stabilization technique twice the corresponding spaces in ( 2 ) are empty . since the stabilization operation does not change the simple homotopy type of a complex , the @xmath13-floer complex before the birth - death is therefore simply homotopy equivalent to the @xmath13-floer complex after the birth - death . due to the assumption that the birth is isolated , there exist a darboux chart @xmath621 , where @xmath622 is a darboux neighborhood centered at @xmath623 such that : 1 . the map @xmath624 is a symplectomorphism @xmath625 , where @xmath626.$ ] 2 . the image of @xmath5 under @xmath624 is modeled by the zero section ; @xmath627 3 . for a horizontal hamiltonian isotopy @xmath628 , the family of lagrangian @xmath629 , is modeled by the graph of the differential of the family of smooth functions @xmath630 defined by @xmath631 where @xmath632 is some non - degenerated quadratic function and @xmath633 then the image of the family under @xmath624 is given by the set @xmath634 the critical points of the family @xmath635 are in correspondence with the intersection points of the family @xmath636 . the family of functions @xmath637 has a unique degenerated critical point at @xmath638 ; since the equation @xmath639 has no solution for @xmath640 and two non - degenerated solutions for @xmath610 . with the previous description we can study the birth - death bifurcation in the lagrangian intersection setting using the classical birth - death bifurcation in a family of morse functions . the quadratic stabilization of a smooth function @xmath641 is the function @xmath642 , where @xmath632 is a non - singular quadratic function on @xmath643 . eliashberg and gromov @xcite used stable morse theory to study lagrangian intersection theory with finite dimensional methods . they remarked that the stabilization process does not change the simple homotopy type of the @xmath644-morse complex . in the present context we will set @xmath645 , so we will study the family of functions @xmath646 . to understand the behavior of the moduli spaces when new intersection points appear we study the stabilized setting defined in analogy with the morse case . the stabilization of the symplectic manifold @xmath647 is defined by @xmath648 the stabilization of a pair of fixed lagrangians @xmath545 , with @xmath649 is defined by @xmath650 and the second lagrangian can be stabilized either in a positive or in a negative way depending on the choice of sign , @xmath651 when required we will denote @xmath652 or @xmath653 to mean a positive or a negative stabilization . in the new setting , after a negative stabilization , the darboux chart @xmath621 is replaced by @xmath654 . the family of lagrangians @xmath629 is now described locally by the stabilized function @xmath655 . thanks to the new dimension we can find a one - parameter family of smooth functions connecting @xmath656 to @xmath657 , close enough to the family @xmath658 but with a birth having a non - zero component on the fiber of the stabilization . to do this we need to have control on the growth of the derivative in the degenerate direction , @xmath659 . we will now change the family of functions @xmath660 which describe locally the birth by the family of functions used by m. sullivan . let @xmath661 be a smooth bump function with the following properties : 1 . @xmath662 is an even function supported on @xmath663 $ ] and non - vanishing on @xmath664 , 2 . there is a unique maximum of 1 at 0 , 3 . @xmath665 has a unique minimum of -2 at @xmath666 , 4 . @xmath667 for @xmath668 . let be @xmath669 . then there exist a darboux chart , still denoted @xmath670 , centered at @xmath623 , such that the family @xmath671 is modeled by the graph of the differential of the family of functions : @xmath672 where @xmath673 is arbitrarily small . the critical points of @xmath637 correspond to intersection points of @xmath674 . notice that if @xmath675 , then @xmath676 when @xmath677 . 1 . if @xmath678 , since @xmath679 , then @xmath680 , so there are no critical points for any parameter @xmath596 in this region . 2 . if @xmath681 , then since @xmath682 has a unique minimum of @xmath683 at @xmath684 then there is a degenerate critical point . 3 . if @xmath685 , then the equation @xmath677 has two solutions for each fixed @xmath596 , so there are two non - degenerate critical points with positive and negative @xmath686-coordinates . the stabilized family of functions @xmath655 can be perturbed as the following theorem asserts : @xcite [ sullivan ] there exists a one - parameter family of functions @xmath687 such that for each @xmath688 : 1 . @xmath689 and @xmath689 , 2 . @xmath690 near the boundary of @xmath691 , 3 . for each @xmath596 , the graph of @xmath692 stays inside @xmath691 . 4 . no deaths of critical points occur and a unique birth occurs arbitrarily close to @xmath693 . 5 . at the moment of the birth @xmath694 . with the help of the previous theorem and the open mapping theorem we prove the following lemma : ( * ? ? ? * theorems 3.7 , 3.12 ) let @xmath569 be a pair of exact lagrangians with cylindrical ends . consider two generic data @xmath695 and @xmath696 defining floer complexes associated to the pair . then there exist a regular homotopy @xmath697 between the generic data such that , if a birth at time @xmath585 occurs , then after stabilizing twice the birth is independent . consider the family of lagrangian @xmath698 defined locally by the functions @xmath699 from theorem [ sullivan ] . this family of lagrangians has a unique birth arbitrarily close to @xmath700 . stabilize in a positive way the new setting . a second application of theorem [ sullivan ] to the new family of functions @xmath699 , implies that there is another family of lagrangians @xmath701 , defined locally by a family of functions @xmath702 , with a unique birth arbitrarily close to @xmath703 . we will show that , for any @xmath704 the moduli spaces of @xmath705-holomorphic strips with only one of the new intersection points involved @xmath706 , are empty : @xmath707 this implies that after stabilizing twice and a double application of theorem [ sullivan ] , the birth is independent . the spaces in ( 11 ) are empty . to see this , let @xmath708 be the local coordinates corresponding to the @xmath709 space added after the first and the second stabilization , where @xmath710 . consider the projection map @xmath711 . this map is @xmath712-holomorphic . we have the following : 1 . the image of @xmath713 under the projection maps @xmath714 is the @xmath715-axis . 2 . assume @xmath716 in theorem [ sullivan ] , then @xmath717 . this implies the the image of @xmath718 under the @xmath714-projection coincides with the image of @xmath719 outside @xmath720 + @xmath721 . thus for @xmath722 the image is the line @xmath723 outside @xmath724 union with some curve inside @xmath724 . for @xmath725 , the image is the line @xmath726 outside @xmath724 union with some curve inside @xmath724 . denote by @xmath727 . suppose there is a strip @xmath728 and consider the map @xmath729 \rightarrow \mathbb{r}^{2}$ ] . this map is holomorphic so it preserves orientations . moreover , due to the boundary conditions , the regions where the image of the strip are mapped ( at least partially ) , by the map @xmath730 are : @xmath731 and @xmath732 . . ] hence , by the open mapping theorem the image of an open subset of the domain @xmath733 $ ] should be an open subset of one of these regions . if this is the case , then the image will contain the unbounded region @xmath734 . this contradicts the fact that the map has finite energy and thus no @xmath705-holomorphic strips from @xmath735 to other intersection point can exist . to see that there are no strips arriving to @xmath735 , we use the same argument but using the second projection @xmath736 . the isotopy given by @xmath737 , where @xmath738 is the hamiltonian generating the family in theorem [ sullivan ] , stabilized twice , satisfies the claim . the following lemma describes the moduli space of strips connecting the newly born intersection points between themselves . * lemma 3.11 ) suppose there exists an @xmath9 and a darboux chart @xmath739 in @xmath740 of the degenerate point @xmath378 , such that for all @xmath741 , there is a symplectomorphism @xmath742 , where @xmath743 is some small interval which contains @xmath744 for some @xmath745 . then , for any @xmath746 , the moduli space @xmath747 contains a unique element . to prove this lemma we show that any strip in @xmath748 has image inside the neighborhood @xmath749 , then we show that the strip should be @xmath90 . suppose @xmath750 , since @xmath751 is the action functional , @xmath752\times [ 0,1]}\tilde{\gamma}^{*}\omega + \int_{0}^{1 } \gamma^{*}h.\ ] ] in the transverse case we set @xmath121 . ] when @xmath753 , and @xmath754 then @xmath755 for @xmath756 small enough . since all the strips lie in the darboux neighborhood , we can describe this situation using the local model . suppose there is a @xmath188-holomorphic strip @xmath757 , consider the map @xmath757 composed with the projection @xmath758 mapping @xmath759 , denoted by @xmath760 , the image of @xmath761 under @xmath762 is a line for @xmath763 . then , applying the open mapping theorem we deduce that @xmath762 is constant for @xmath763 and @xmath764 . the floer complex associated to a pair of lagrangian submanifolds @xmath545 and the regular data @xmath132 coincide with the complex associated to the stabilized data . there are no new intersection points , and by a similar argument to the one used before with involving the open mapping theorem , the moduli spaces are the same . for a detailed proof see ( * ? ? ? * theorem 3.9 ) . the immediate consequence of the previous lemmas and the remark is the following description of the floer complex after a birth occurs , for a regular homotopy : @xmath765 where @xmath766x^{+}_{\lambda + \epsilon}.$ ] for definition of @xmath120 , see equation [ def - teta ] . let @xmath767 , @xmath768 be two families of non - degenerate intersection points of @xmath769 . let @xmath90 be a handle - slide from @xmath770 to @xmath771 occurring at @xmath585 . for any two families of non - degenerate intersection points @xmath772 , the @xmath773-dimensional one - parameter moduli space is given by : @xmath774 to describe the change in whitehead torsion of the complex after a handle - slide bifurcation , it is required to understand the boundary components of the compactified one - parameter moduli spaces @xmath775 and @xmath776 , for any one - parameter family of intersection points @xmath777 with @xmath778 or @xmath779 , @xmath780 . the gromov compactness theorem states that the moduli spaces can be compactified by broken trajectories . on the other hand , a gluing theorem shows that each broken trajectory is a boundary component of the moduli space . these two results together imply the following relations : @xmath781}(x^{\lambda}_{- } , y^{\lambda } ) & = -\widehat{\mathcal{m}}_{\lambda_{0}-\epsilon}^{0}(x^{\lambda_{0}}_{- } , y^{\lambda_{0 } } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}+\epsilon}^{0}(x^{1}_{- } , y^{1 } ) \cup -\{u\}\times\widehat{\mathcal{m}}_{\lambda_{0}}^{0}(x^{\lambda_{0}}_{+ } , y^{\lambda_{0}})\\ \partial \overline{\mathcal{\hat{m}}}^{0}_{[\lambda_{0}-\epsilon , \lambda_{0}+ \epsilon]}(x^{\lambda } , x^{\lambda}_{+ } ) & = -\widehat{\mathcal{m}}_{\lambda_{0}-\epsilon}^{0}(x^{\lambda_{0 } } , x^{\lambda_{0}}_{+ } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}+\epsilon}^{0}(x^{1 } , x^{1}_{+ } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}}^{0}(x^{\lambda_{0 } } , x^{\lambda_{0}}_{-})\times\{u\},\end{aligned}\ ] ] and out of this we can relate the whitehead torsion of the floer complexes before and after a handle - slide . to show ( 12 ) and ( 13 ) we need to prove a gluing theorem . the standard gluing theorem concerns the gluing of non - degenerate trajectories when the linearization of the cauchy - riemann operator at each trajectory is surjective . if a degenerate orbit appears , the linearization of the cauchy - riemann operator at the degenerate trajectory has a one - dimensional cokernel . in this situation we state the gluing theorem as follows , [ gluing ] suppose there is a handle - slide @xmath782 ( @xmath783 ) , for some @xmath784 $ ] . then for every intersection point @xmath785 with @xmath786 , there is @xmath787 and a continuous map : @xmath788 such that for every @xmath789 , if @xmath790 is a sequence converging to @xmath791 , then @xmath792 for @xmath102 big enough . an analogous statement holds for @xmath793 , when @xmath794 satisfies @xmath795 for any @xmath780 . a stronger version of theorem 4.4 appears already as proposition 4.2 in floer s article @xcite , where he also suggested the proof . lee gives a proof of theorem 4.4 in ( * ? ? ? * theorem 7.5),for the floer theory of hamiltonian periodic orbits . we will present lee s proof adapted to the lagrangian case . we use the naturality of floer s equation , which leads to the identification of the moduli space of floer strips with the moduli space of floer half - tubes . to a given @xmath796 associate an orbit of the hamiltonian vector field , @xmath797 . in the same way a floer s strip is transformed into a floer s half - tube : @xmath798 where @xmath799 satisfies the floer perturbed equation : @xmath800 an equation of the form : @xmath801 with the above identification , we rewrite theorem [ gluing ] as : suppose there is a handle - slide @xmath802 ( @xmath803 ) , for some @xmath784 $ ] . then for every orbit @xmath804 with @xmath805 , there is @xmath787 and a continuous map : @xmath806 such that for every @xmath807 , if @xmath808 is a sequence converging to @xmath791 , then @xmath809 for @xmath102 big enough . the analogous statement holds for @xmath810 , where @xmath811 satisfies @xmath812 . the theorem is proven in three steps . * first step : _ construction of a pre - gluing map . _ a point @xmath813 is a solution of the equation ( 14 ) at time @xmath596 . in this step an approximated solution @xmath814 depending on the data @xmath815 is defined . the map assigning to each @xmath816 a pair @xmath814 is call the pre - gluing map . + the map @xmath817 does not satisfy equation ( 14 ) , but it is close to a real solution since @xmath818 is sufficiently small . * second step : _ construction of @xmath819 . _ in this step we use the newton - picard method stated in lemma [ contraction ] , to construct @xmath819 . the newton - picard method is applied to the operator @xmath820 . starting with a pre - glued solution @xmath814 ; the method gives us a honest solution . in order to apply this method the following is needed : 1 . the existence of a uniformly bounded right inverse of the linearization of the operator @xmath821 denote by @xmath822 the linearization of @xmath823 at @xmath814 . + in this step it is shown that @xmath822 has a right inverse , denoted by @xmath824 , and that this right inverse is uniformly bounded independently of @xmath816 . 2 . obtaining a quadratic bound on the non - linear part of @xmath823 . + near @xmath814 , the operator @xmath823 can be decomposed into linear part and a non - linear part . let @xmath825 , @xmath826 if @xmath827 , then a solution of @xmath828 is obtained as the fixed point of @xmath829 the existence of such a fixed point is guaranteed by the contraction mapping theorem which implies the following lemma : + [ contraction ] let @xmath830 be the upper bound on @xmath831 as above , and suppose that there is a @xmath56-independent constant @xmath832 such that @xmath833 + then there exists a unique @xmath834 with @xmath835 solving ( 3.2.7 ) . moreover , the solution @xmath834 varies smoothly with @xmath56 , and @xmath836 + the gluing map is defined by @xmath837 . * third step : _ show the `` surjectivity '' of the gluing map : _ in this step we show that if @xmath838 is a sequence converging to @xmath791 , then for all @xmath102 large enough , we have @xmath809 . let @xmath839 and @xmath840 , with @xmath841 for each @xmath780 ; @xmath842 , \tilde{m}),\hspace{0.2 cm } \lim\limits_{s \rightarrow \mp \infty } u(s , t ) = \gamma_{\mp}^{\lambda}(t)\}.\ ] ] for a vector bundle @xmath843 $ ] denote by @xmath844 the space of @xmath845-sections @xmath846 \rightarrow e$ ] . for @xmath847 , \tilde{m})$ ] with boundary condition on a lagrangian submanifold @xmath848 , let @xmath849 we now proceed to the proof : * step 1 . * construction of pre - gluing map . let @xmath850 and @xmath851 $ ] be two smooth functions , @xmath852 and @xmath853 pick @xmath854 large enough and let @xmath855 be the exponential map associated to a riemmanian metric @xmath856 for which @xmath5 is totally geodesic . suppose that for @xmath857 we have @xmath858 . the pre - gluing map is defined as follows : @xmath859 where @xmath860\rightarrow m$ ] is given by @xmath861 the map @xmath862 , has the following properties : * @xmath863 in @xmath864 , * @xmath865 in @xmath866 * @xmath867 in @xmath868 in the same way we define the linearization of the pre - gluing map for @xmath56 fixed @xmath869 as follows : @xmath870 where @xmath24 varies in the same intervals as in the non - linear case and @xmath871 the section @xmath872 has the following properties : * @xmath873 in @xmath874 , * @xmath875 for @xmath876.$ ] the pregluing map satisfies @xmath877 . * step 2 . * construction of @xmath819 . we first show the existence of a uniformly bounded right inverse for the linearization of the operator @xmath823 ( [ operatopr ] ) at the point @xmath814 . in order to do this we recall some definitions and standard results about the operator @xmath823 . for @xmath878 let @xmath879 , \tilde{m } ) \hspace{0.2cm}|\hspace{0.2 cm } w(s , i ) \in w \hspace{0.2 cm } i=0,1 \}.\ ] ] for a fixed @xmath880 , the space @xmath881 is an infinite dimensional banach manifold ( see ( * ? ? ? * theorem 3 ) ) . consider the vector bundle @xmath882 , where the base space is defined by @xmath883 and the fiber is given by @xmath884 . the perturbed cauchy - riemann operator defines a section of the following bundle , @xmath885 given by @xmath886 the space @xmath887 is a banach manifold with local charts modeled by @xmath888 and local coordinates given by exponential map : @xmath889 for two large enough positive constants @xmath890 . with the previous definitions we have that @xmath891 . then , in a neighborhood of @xmath892 the section has the local expression : @xmath893 to apply lemma [ contraction ] to the operator @xmath894 , we first have to show the existence of a uniformly bounded right inverse for the linearization of the operator @xmath895 at @xmath896 . denoting the linearized operator by @xmath822 , this operator depends on the choice of a connection and on the constants @xmath890 used in the definition of the local coordinates . the linearized operator has the following expression : @xmath897 the first term is @xmath898 , where @xmath899 is a zero - order operator such that @xmath900 . the operators @xmath901 are symmetric operators on the solutions of the perturbed cauchy - riemann equation . the second term @xmath902 is the component on the @xmath596-direction and is a cut - off version of : @xmath903 the operator @xmath822 is a fredholm operator ( see theorem 4 @xcite ) , with index @xmath904 . to see that the operator @xmath822 has a right inverse , observe that @xmath905 ( a proof of this can be found in ( * ? ? ? * lemma 1.2.4 ) ) , then , @xmath906 . this combined with the computation of the index implies that the operator @xmath822 is surjective , since it is fredholm then it has a right inverse . denote its right inverse by @xmath907 . to see that the operator @xmath907 is bounded , we have to show that there exist @xmath908 such that , for @xmath909,@xmath910 we proceed by contradiction . assume that the inequality above is not true . then there exist a sequence @xmath911 with @xmath912 and there exist @xmath913 such that @xmath914 let @xmath915 $ ] be a smooth bump function such that @xmath916 if @xmath917 and @xmath918 if @xmath919 . consider @xmath920 and the operator @xmath921 where @xmath922 and @xmath923 is the linearization at @xmath145 of the hamiltonian flow equation @xmath924 the following estimates show that @xmath925 converges to zero in @xmath926-norm : @xmath927 where @xmath68 is the upper bound of @xmath928 on @xmath97 and @xmath929 \times [ 0,1])$ ] . notice that @xmath930 , and since @xmath931 , this implies ( 3.2.20 ) , for some @xmath932 that goes to @xmath118 when @xmath102 goes to infinity . on the other hand @xmath933 in @xmath864 , which implies that @xmath934 goes to zero in @xmath926-norm . notice now that @xmath935 is a fredholm operator , hence there exist a subsequence of @xmath936 that converge to some element @xmath937 with @xmath938 bounded . consider the integral kernel @xmath939 on @xmath97 , the integral kernel of @xmath935 is the function @xmath940 associated to the operator @xmath941 such that @xmath942 if @xmath943 . ] of the operator @xmath944 . since the operator @xmath945 is invertible and @xmath946-self - adjoint , the integral kernel of the operator @xmath944 is given by the following expression : @xmath947 where @xmath948 are the orthogonal projections and @xmath949 are the restrictions of @xmath923 to the subspace of @xmath950 generated by the positive or negative eigenvectors ( this computation can be found in ( * ? ? ? * a ) ) . applying the operator @xmath939 on both sides of the equation @xmath951 we obtain @xmath952 however , @xmath953 is constant in @xmath24 and @xmath954 has exponential decay on @xmath955 . this implies the exponential decay of @xmath956 when @xmath957 . therefore , the only pair that satisfies equation ( [ equ1 ] ) is @xmath958 . this implies that @xmath959 . consider the sequences @xmath960 and @xmath961 . after extracting a subsequence , we can show that these sequences converge to zero in @xmath962 . the proof of this is completely analogous to the standard case . for a detailed proof , the reader can consult ( * ? ? ? * lemma 9.4.10 - 9.4.12 ) . finally , we have the following estimate : @xmath963 where the terms on the right tends to zero , giving a contradiction . thus @xmath907 has a uniform bound . to apply lemma [ contraction ] , we need to obtain a quadratic bound on the non - linear part of @xmath895 . this bound follows from the following estimates : @xmath964 } \vert ( dn)_{(\mu , \xi)}\vert \vert ( \mu_{1 } , \xi_{1 } ) - ( \mu_{2 } , \xi_{2})\vert_{1,p } \\ & \leq k \sup \limits_{(\mu , \xi ) \in [ ( \mu_{1 } , \xi_{1 } ) , ( \mu_{2 } , \xi_{2 } ) ] } \vert ( \mu , \xi ) \vert_{1,p } \vert ( \mu_{1 } , \xi_{1 } ) - ( \mu_{2 } , \xi_{2})\vert_{1,p } \\ & \leq k ( \vert ( \mu_{1 } , \xi_{1})\vert_{1,p } + \vert ( \mu_{2 } , \xi_{2})\vert_{1,p } ) \vert ( \mu_{1 } , \xi_{1 } ) - ( \mu_{2 } , \xi_{2})\vert_{1,p}. \end{aligned}\ ] ] inequality ( 3.2.29 ) follows from a long computation that can be found for example in ( * ? ? ? * theorem 3.a ) or in the following lemma : @xcite let @xmath965 . there exist a constant @xmath241 such that , for all @xmath909 and all @xmath966 with @xmath967 , we have @xmath968 at this point , lemma [ contraction ] applies , and we obtain that there exist a unique @xmath969 solving equation ( 17 ) : @xmath970 . the gluing map is defined as follows : @xmath971 here , @xmath972 is the projection to the unparametrized moduli space . once defined the gluing map , we arrive at the final step . * step 3 . * in this step we show the so called `` surjectivity '' of the gluing map . the surjectivity in this context means that given a sequence @xmath838 converging to @xmath791 , there is @xmath973 such that for all @xmath974 we have @xmath809 . equivalently , we will show that any trajectory in the moduli space @xmath975 , in a small neighborhood of the the broken trajectory @xmath976 denoted by @xmath977 , is in the image of the gluing map . to do so , we see that for any @xmath9 small enough there is a @xmath978 such that the neighborhood of a broken trajectory @xmath976 , @xmath979 satisfies @xmath980 where @xmath981 is a neighborhood of the pre - glued trajectories defined as follows : @xmath982 we first show that if @xmath983 , then @xmath984 for some choice of @xmath985 and @xmath986 in a small chart centered at @xmath987 satisfying @xmath988 ( for the norm induced by the riemannian metric ) . to complete the proof we have to verify that @xmath989 , which implies that @xmath990 . let us choose @xmath991 , such that @xmath992 and @xmath993 ( such a @xmath56 can be found by re - parametrization ) . let be @xmath230 with the property that outside @xmath994 \times [ 0 , 1]$ ] the value of @xmath995 is in a small ball @xmath996 for @xmath997 . since @xmath998 is also close to @xmath999 , we can suppose that @xmath756 is smaller than the injectivity radius of @xmath855 for a big enough @xmath1000 . hence we can write @xmath1001 for some @xmath1002 with @xmath1003 for the norm induced by the metric . observe now that @xmath1004 . to see that @xmath1005 , decompose @xmath1006 $ ] , the domain of @xmath1002 , in the following sub - domains : @xmath1007 $ ] , @xmath1008 $ ] , @xmath1009 $ ] , @xmath1010 $ ] and @xmath1011 $ ] . the weak compactness guarantees that in @xmath1012 , @xmath1013 , where @xmath1014 is a constant independent on @xmath1000 and @xmath56 . the exponential decay of @xmath817 and @xmath757 to @xmath1015 implies that @xmath1016 on @xmath1017 . finally on @xmath1018 , there is @xmath1019 such that @xmath1020 since @xmath757 satisfies the perturbed cauchy - riemann equation , @xmath1021 satisfies the following equation : @xmath1022 where @xmath1023 is the non - linear term left . applying the integral kernel to the equation below we obtain that : @xmath1024 where @xmath1025 on @xmath1018 . by definition , @xmath1026 . moreover , using the exponential growth / decay of @xmath1027 , we can estimate : @xmath1028 and therefore @xmath1029 the exponential decay of @xmath939 implies a uniform bound on @xmath1030 and @xmath1031 . on the other hand , from the exponential decay of @xmath817 to @xmath145 over @xmath1018 , we obtain the estimate @xmath1032 , for some @xmath1033 , and thus since similar estimates follow for @xmath1034 and @xmath1035 , we obtain that @xmath1036 summing up the estimates on each region of @xmath1037 , and putting @xmath1038 , we obtain the bound @xmath1039 the expressions , @xmath1040}(x^{\lambda}_{- } , y^{\lambda } ) & = -\widehat{\mathcal{m}}_{\lambda_{0}-\epsilon}^{1}(x^{\lambda_{0}}_{- } , y^{\lambda_{0 } } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}+\epsilon}^{1}(x^{1}_{- } , y^{1 } ) \cup -\{u\}\times\widehat{\mathcal{m}}_{\lambda_{0}}^{1}(x^{\lambda_{0}}_{+ } , y^{\lambda_{0}})\\ \partial \bar{\widehat{\mathcal{m}}}^{1}_{[\lambda_{0}-\epsilon , \lambda_{0}+ \epsilon]}(x^{\lambda } , x^{\lambda}_{+ } ) & = -\widehat{\mathcal{m}}_{\lambda_{0}-\epsilon}^{1}(x^{\lambda_{0 } } , x^{\lambda_{0}}_{+ } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}+\epsilon}^{1}(x^{1 } , x^{1}_{+ } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}}^{1}(x^{\lambda_{0 } } , x^{\lambda_{0}}_{-})\times\{u\},\end{aligned}\ ] ] follow from the previous gluing theorem and the choice of orientations . it also implies the corollary [ colh - s ] below . to state it , consider the following linear map between the floer complex before and after the handle - slide : @xmath1041\delta^{x^{\lambda_{0 } - \epsilon}}_{x_{-}^{\lambda_{0 } - \epsilon } } x_{+}^{\lambda_{0 } + \epsilon}.\end{aligned}\ ] ] here , sign@xmath119 , will be object of the next section were we will discus orientations and we will fix the conventions used in the present work . ] is a sign coming from the conventions on orientation . [ colh - s ] the map @xmath261 defined before is a chain map . thus , the floer complex after a handle - slide @xmath1042 occurs is described by : + @xmath1043 to simplify the notation , let us write @xmath1044 . notice that the only difference on the differential occurs when computing @xmath1045 or @xmath1046 on an intersection point @xmath1047 with @xmath1048 . from the gluing theorem we have the relation @xmath1049d_{-}x_{+}^{\lambda_{0 } - \epsilon } = 0 $ ] , which clearly implies that @xmath1050 . for the second case , let @xmath1047 be such that @xmath1048 . the gluing theorem implies the relation @xmath1051x_{+}^{\lambda_{0 } + \epsilon } = 0 $ ] , thus @xmath1052 . this completes the description of the bifurcations along the family @xmath1053 . we will use the following lemma . let @xmath498 be a @xmath13-complex . 1 . if @xmath1054 is a trivial acyclic @xmath13-complex @xmath1055 , where @xmath1056 and the corresponding preferred bases are isomorphic . then , @xmath1057 . the chain map @xmath1058 , @xmath1059 satisfies @xmath1060 . finally we have all the elements to prove the results of the section . choose a regular homotopy @xmath1061 joining @xmath1062 with @xmath1063 where @xmath1064 is a horizontal isotopy . such a homotopy exists from proposition 4.1 . assume that a birth arrives at @xmath585 . since the birth is isolated , we can restrict to a sub - interval @xmath1065 $ ] where the only bifurcation is the birth . we now apply lemma 4.1 to find a regular homotopy that makes the birth independent after stabilizing twice . from remark [ birth - death ] and the lemma 4.5 , the whitehead torsion does not change after a birth . we treat in the same way a death bifurcation . suppose now that there is a handle - slide at @xmath585 . the gluing theorem and corollary 4.1 and lemma 4.5 imply that the whitehead torsion does not change . finally , any intersection point @xmath127 generates a path @xmath1066 along the isotopy . then , the matrix @xmath1067 associated to the complex @xmath1068 may differ by some factors on @xmath13 , from the matrix @xmath1069 associated to the floer complex @xmath1070 . since for any intersection point , @xmath1071 is a loop . this difference can be expressed with a series of elementary operations . this implies that the whitehead torsion of the two complexes coincide . by corollary 2.1 , an exact - orientable- spin lagrangian cobordism @xmath0 , such that the map @xmath1072 induced by the inclusion @xmath31 is an isomorphism for @xmath32 , is an h - cobordism . thus , it has a well defined whitehead torsion . the whitehead torsion satisfies : @xmath1073 the equality ( 51 ) , follows from theorem 3.3 and remark 3.1 . the equality ( 52 ) follows from the proof of proposition 2.4 . for any morse function without critical points on a collar neighborhood of the boundary , the whitehead torsion of the cobordism coincides with the whitehead torsion of the @xmath13-morse complex . here , @xmath1074 is a pair composed by a morse function and @xmath1075 is the riemannian metric induced by an almost complex structure @xmath1076 as defined in section [ chapter floer homology ] . then , we can compute the whitehead torsion of the cobordism @xmath0 , using the @xmath13-morse complex @xmath1077 . since for small hamiltonian function , @xmath1078 as in proposition 2.4 , the morse complex and the floer complex coincide , we obtain ( 52 ) . the equality ( 53 ) follows from the theorem [ theorem invariance ] . since the torsion of the floer complex is independent on the choice of horizontal isotopy , and there is a horizontal isotopy that displaces @xmath1079 from @xmath5 ( where @xmath1080 ) , for such a pair @xmath1081 the floer complex is trivial so the whitehead torsion of this complex vanishes . let @xmath61 be a pair of lagrangians intersecting transversely and @xmath103 be a generic pair composed by a compactly supported hamiltonian function and a time dependent almost complex structure @xmath188 . for @xmath1082 , @xmath1083 , m ) \hspace{0.2cm}\mid \hspace{0.2 cm } \substack { u(s , i ) \in l_{i},\hspace{0.2 cm } \bar{\partial}_{\mathbf{j}}(u)= 0\\ \lim \limits_{s \to \mp \infty } u(s , t ) = x / y , \hspace{0.2 cm } e(u ) < \infty } \},\ ] ] @xmath1084 , m ) \hspace{0.2cm}\mid \hspace{0.2 cm } \substack{u(s , 0 ) \in l_{0},\hspace{0.2 cm } u(s,1 ) \in \varphi_{\beta(s)}(l_{1 } ) , \\ \bar{\partial}_{\mathbf{j}}(u)= 0\\ \lim \limits_{s \to \mp \infty } u(s , t ) = x / y , \hspace{0.2 cm } e(u ) < \infty } \}.\ ] ] for @xmath1085 , where @xmath1086 is the space of paths connecting @xmath6 with @xmath7 , that are orbits of the hamiltonian flow @xmath77 , @xmath1087,m ) \hspace*{0.2 cm } | \hspace*{0.2 cm } \substack{u(s , i ) \in l_{0 } , \\ \overline{\partial}_{\mathbf{j}}u = - \nabla h(u , t),\\ \lim \limits_{s\to \pm\infty}u(s , t ) = \gamma_{\pm}(t ) , e(u ) < \infty \\ } \}.\ ] ] let @xmath1088 be a one parameter family of regular data at each time @xmath596 . then there are families of orbits @xmath1089 , @xmath1090 the propose of this section is to fix the conventions used to orient the previous moduli spaces . we will also recall some known results concerning the orientation problem for these moduli spaces , following @xcite and @xcite closely . let @xmath1091 be banach spaces . to any fredholm operator @xmath1092 we can associate a one - dimensional real vector space called the _ line determinant _ : @xmath1093 given a family of fredholm operators @xmath1094 , the vector bundle induced by taking the determinant line at each parameter @xmath1095 is a line bundle , called the _ determinant line bundle_. to a moduli space of @xmath1096-holomorphic curves @xmath1097 one naturally associates a family of fredholm operators , giving rise to a determinant line bundle on it . the problem of orientability of a moduli space translates to the study of its associated determinant line bundle . an exact sequence of finite - dimensional vector spaces @xmath1107 induces an isomorphism between @xmath1108 by writing @xmath1109 for appropriated oriented subspaces @xmath1110 where @xmath1111 restrict to isomorphisms . let @xmath5 be an oriented manifold with boundary . denote by @xmath1114 its orientation and by @xmath1115 an exterior pointing vector on @xmath410 . then @xmath410 is oriented by the following convention : @xmath1116 let @xmath1098 be a fredholm operator with oriented kernel @xmath1117 and cokernel @xmath1118 . given a stabilization of it @xmath1119 , we orient its determinant line @xmath1120 by the induced orientations from the following exact sequence : @xmath1121 in particular , from the last choice we can conclude that the orientation of the moduli spaces @xmath1129 ( see [ j - hom strip ] ) , appearing in section [ chapter floer homology ] , section 1.1 , are compatible with the gluing theorem since in this case the operators are surjective so there is no difference of sign . the moduli space of interest here is @xmath1130 ( see [ j - hom strip ] ) , the space of floer strips with boundary conditions along a pair of lagrangians @xmath61 . in @xcite , a sufficient condition for the orientability of this space the condition imposed is that the pair @xmath61 has to be ( relatively ) spin . * definition 2.1.9 ) let @xmath1037 be an oriented compact surface with boundary . a complex bundle pair @xmath1132 is a complex vector bundle @xmath1133 with a maximal totally real bundle @xmath1134 , such that @xmath1135 . we now recall the results about the orientability of moduli spaces . the following proposition is used to assign a canonical orientation at any determinant line associated to a point in the moduli space . the orientability of the moduli spaces of floer strips is proved in ( * ? ? ? * theorem 8.1.14 ) . the proof uses proposition [ propfo ] combined with a gluing argument . additional choices of spin structures and orientations of auxiliary determinant lines are also required . here , we will present the proof of an analogous statement ( [ orientability ] ) , for the moduli spaces of half - tubes @xmath1138 ( defined in [ jh - tube ] ) with boundary condition on a fixed lagrangian . the idea of the proof is the following . each orbit @xmath1139 satisfies @xmath1140 = i d \in \pi_{1}(m , l)$ ] . let @xmath1141 \times \mathbb{r}$ ] be the half disk . there exist a map @xmath1142 such that @xmath1143 . this map is a capping for @xmath811 . a gluing argument applied to an appropriate one - parameter family of operators relating the operator @xmath1150 with the operator @xmath1151 ( the linearization of the operator @xmath1152 at @xmath90 ) induces a canonical orientation on @xmath1153 . after a canonical orientation is given to each @xmath1154 , the spin condition guarantees a global orientation on @xmath1138 . let @xmath878 , and consider the following two operators : @xmath1171 which is the cauchy - riemann operator with lagrangian boundary condition and the operator @xmath1172 is the linearization of the operator @xmath1173 defined as follows . let @xmath1174 $ ] be a smooth function such that @xmath1175 the operator is given by @xmath1176 for @xmath1177 and a choice of hermitian connection @xmath1178 for which @xmath188 is parallel , the linearization is given by : @xmath1179 by proposition [ propfo ] , the determinant line @xmath1180 has a canonical orientation ( since its the sum of the cauchy - riemann operator with a zero - order operator ) , induced from a trivialization of @xmath1181 . the gluing theorem for fredholm operators ( see for example @xcite ) implies the isomorphism between the determinant lines : @xmath1182 from [ detiso ] , it follows that a choice of orientation on @xmath1183 determines an orientation on @xmath1184 . it remains to show that the previous choices lead to a trivial determinant line bundle @xmath1185 to see this , let @xmath1186 be in the same connected component , orient @xmath1187 using the trivialization that comes from the spin structure on @xmath48 and a choice of orientation on @xmath1183 . consider a path @xmath1188 joining @xmath1189 to @xmath1190 . since @xmath1191 is trivial , we can transport the orientation to @xmath1192 . the orientation induced on @xmath1192 does not depend on the choice of path since for any other path @xmath1193 , we can consider the loop given by , @xmath1194 since @xmath48 is spin , for any @xmath1195 , the spin structure gives a trivialization of @xmath1196 which extends to a trivialization of @xmath1197 , since this trivialization is unique up to homotopy the induced orientation is also unique . from the discussion above we can conclude the following theorem : [ orientability ] suppose that a lagrangian @xmath48 is spin . then for any @xmath1198 , the moduli space @xmath1199 ( see [ jh - tube ] ) of connecting orbits in lagrangian floer theory @xmath1200 , is orientable . furthermore , orientations on @xmath1201 and the spin structure canonically determine an orientation on @xmath1199 . theorem [ orientability ] is based on theorem @xcite , wich establishes conditions for the orientability of the moduli space @xmath1202 ( defined in expression [ j - hom strip ] ) , of floer strips connecting intersection points , with boundary condition in a pair of lagrangians @xmath61 . theorem [ orientability ] is less general , since it refers to moduli spaces of floer half tubes ( defined in [ jh - tube ] ) , connecting trajectories of the hamiltonian vector field with boundary condition on a fixed lagrangian , these are the ones used in section [ chapitre whitehead ] for the bifurcation analysis . the proof of theorem 8.1.14 in @xcite is analogous to the one of the theorem above . some differences are the following : instead of choosing capping disks associated to each trajectory @xmath142 , a path @xmath1203 , joining @xmath1204 with @xmath1205 is associate to each intersection point @xmath1206 in a canonical way using the spin structures on @xmath4 . now , for each floer strip @xmath1207 , consider the loop @xmath1208 . to orient @xmath1214 , we use virtual operators defined on half - disks with boundary conditions involving the paths @xmath1215 and @xmath1216 . from a gluing argument on these bundles we obtain an expression analogous to the one in [ detiso ] . finally , we use the spin structures to transport the orientation on the connected component of @xmath90 . when moving boundary conditions are present , ( * ? ? ? * theorem 8.1.14 ) still applies : we can transform a strip with moving boundary conditions into a strip with fixed boundary conditions using the map @xmath1217 , defined in section 1.1.2.4 . the new strip satisfies a perturbed @xmath1218-type equation . the linearization of it is a fredholm operator . the spin structure induces an orientation on the determinant line bundle of the moduli space composed by strips with fixed boundary condition . since this moduli space is diffeomorphic to the space with moving boundary condition , the diffeomorphism induces an orientation on it . [ [ orientability - of - the - moduli - spaces - mathcalm_lambdagamma_-lambda - gamma_lambdaj - parm ] ] orientability of the moduli spaces @xmath1219([j - parm ] ) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ let @xmath575 $ ] and consider a generic homotopy @xmath576 joining two generic data @xmath577 and @xmath578 and assume that for each @xmath596 the data @xmath1220 is regular . the spin structure on @xmath48 is transported along the homotopy @xmath1221 . assume that @xmath1222 are two one - parameter families of non - degenerate orbits . let @xmath1223 . following the same argument as the one given in the proof of theorem [ orientability ] , the spin structure grantees that the orientation can be transported on a connected component of the base @xmath1235 . then , @xmath1235 has a canonical orientation up to the choices . determinant lines also induce orientations on spaces of flow lines connecting critical points of a morse function . the orientations induce by the determinant bundles coincide with the classical orientation of these spaces given by the intersection of stable and unstable manifold ( under some choices ) . this construction appears in appendix b @xcite . in particular , this implies that in the exact case treated in section 1.2.3 , the differential of the floer complex and morse complex agree . since the moduli spaces [ j - hom strip ] are identified with the spaces of flow lines connecting critical points , and the orientations of both spaces can be defined by determinant bundles . let @xmath1236 be a 0-dimensional moduli space of unparametrized trajectories . let @xmath90 be any representative of @xmath1237 in @xmath1238 . then sign@xmath1239 = sign@xmath119 . is defined by the relation @xmath1240 , where the last space has a canonical orientation induced by the spin structure and capping . for @xmath1241 , we set @xmath1242 . here @xmath1243 is defined in section [ chapter floer homology ] , ( 7)subsection 2.1.4 , and @xmath1244 is the linearization of the @xmath1245-type operator defined by the equation satisfied by @xmath1243 . in this subsection we show that the signs of the boundaries of the compactified one - parameter moduli spaces @xmath1249 satisfy : @xmath1040}(x^{\lambda}_{- } , y^{\lambda } ) & = -\widehat{\mathcal{m}}_{\lambda_{0}-\epsilon}^{1}(x^{\lambda_{0}}_{- } , y^{\lambda_{0 } } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}+\epsilon}^{1}(x^{1}_{- } , y^{1 } ) \cup -\{u\}\times\widehat{\mathcal{m}}_{\lambda_{0}}^{1}(x^{\lambda_{0}}_{+ } , y^{\lambda_{0}})\\ \partial \bar{\widehat{\mathcal{m}}}^{1}_{[\lambda_{0}-\epsilon , \lambda_{0}+ \epsilon]}(x^{\lambda } , x^{\lambda}_{+ } ) & = -\widehat{\mathcal{m}}_{\lambda_{0}-\epsilon}^{1}(x^{\lambda_{0 } } , x^{\lambda_{0}}_{+ } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}+\epsilon}^{1}(x^{1 } , x^{1}_{+ } ) \cup \widehat{\mathcal{m}}_{\lambda_{0}}^{1}(x^{\lambda_{0 } } , x^{\lambda_{0}}_{-})\times\{u\}.\end{aligned}\ ] ] a similar statement appears as ( * ? ? ? * section 7.3.2 ) . see subsection 4.2 , section [ chapitre whitehead ] to recall the definitions . the determinant lines @xmath1253 and @xmath1254 are oriented as theorem [ orientability ] explains . then we have the following isomorphisms : @xmath1255 the isomorphism ( 65 ) follows from the conventions for the gluing map in subsection 5.2.4 . we conclude that sign@xmath119sign@xmath1256 = -sign@xmath1257 . notice that the orientation of @xmath1258 is given by @xmath1259 . then , the base @xmath1260 is a compatible base when @xmath1261 . from this we can conclude that @xmath1262 . in the same way we deduce ( 62 ) . this concludes the verification . auroux d. a beginner s introduction to fukaya categories . in frdric bourgeois , vincent colin , and andrs stipsicz , editors , _ contact and symplectic topology _ , volume 26 of _ bolyai society mathematical studies _ , pages 85136 . springer international publishing , 2014 .
we show that under some topological assumptions , an exact lagrangian cobordism @xmath0 of dimension @xmath1 is a lagrangian pseudo - isotopy . this result is a weaker form of a conjecture proposed by biran and cornea , which states that any exact lagrangian cobordism is hamiltonian isotopic to a lagrangian suspension .
advanced - generation interferometric gravitational wave detectors , such as advanced ligo @xcite , advanced virgo @xcite and kagra @xcite are currently being commissioned . their sensitivity is expected to surpass that achieved by first generation instruments by almost an order of magnitude in the high frequency region . to achieve this , very high circulating power levels ( 0.5 - 1 mw ) will be stored within the fabry - perot arm cavities . at these power levels , even low levels of optical absorption can lead to significant thermoelastic distortion of optical surfaces and unacceptable levels of wavefront distortion @xcite , resulting in reduced circulating power and a reduction in the efficiency of the detector signal readout . thermally actuated compensation systems will be thus used to ameliorate the wavefront distortion . however , the thermal time constants for the absorption - induced distortion and the compensation are long , typically 12 hours , and thus incorporating predictive modeling in the control systems may prove essential . the response of a linear elastic system to heating is described by the theory of thermo - elasticity and its applications to highly symmetric , idealized systems are described in many books ( see @xcite for example ) . it has also been used to develop analytic expressions for less idealized optical systems @xcite . the expressions developed by hello and vinet @xcite are relevant to the work described here , but apply only to cylindrical isotropic mirrors heated by coaxial laser beams . more complicated systems , which incorporate asymmetric heating or anisotropic elasticity , can be investigated using finite - element numerical models that apply the equations of thermo - elasticity on a three - dimensional spatial mesh . for dynamic systems , the thermoelastic equations must be solved at each epoch , requiring computational times that can run to many days . this approach would be untenable for use in predictive feed - forward actuation to control systems . in such cases , the solution of the scalar problem to determine the temperature profile throughout the optic can be solved rapidly ; the time consuming part is solving the tensor - based elasticity problem to convert the thermal profile into an elastic distortion . the betti - maxwell theorem of elastodynamic reciprocity @xcite provides an alternative approach to using finite - element methods ( fem ) to solve the tensor part of the thermoelastic distortion . it has previously been used to investigate the excitation of rayleigh - lamb elastic waves in a metal plate due to heating produced by a line - focused pulsed laser beam assuming that the heating is confined to the surface of the plate and it has infinite lateral extent @xcite . in the context of gravitational wave detection , it has been used to compute the interferometer s response to creep events in the fibers that suspend the optics @xcite . we extend its use to predict thermoelastic distortion of an optic of finite size with asymmetric heating . we describe here how elastodynamic reciprocity and fem can be combined to provide accurate predictions of thermoelastic surface distortion more quickly than using fem alone . in summary , fem is used to determine the response of the optic to a set of orthonormal tractions , or pressures a computationally expensive calculation that is performed once for an optic . then , using reciprocity , the distortion due to the instantaneous temperature profile in the optic is calculated using a sum of scalar volume integrals that incorporate these responses . the computational cost of this step is much less than that of a full elastostatic fem evaluation . additionally , it is amenable to parallelization , which would further reduce the computational time . the layout of the rest of the paper is as follows : in section ii we introduce the betti - maxwell theorem of elastodynamics and show how it can be used to determine the surface distortion by careful choice of a suitable ` auxiliary ' elastic system . we demonstrate its application by calculating the distortion of the end face of a cylindrical optic that is heated by a gaussian heat flux that is ( a ) coaxial with and ( b ) laterally displaced from the axis . the approach and model are described in sections iii and iv . finally , the resulting surface distortions are presented in section v and compared with the results of elastostatic fem calculations . computation times for these two approaches are compared in section vi the betti - maxwell reciprocity theorem for elastodynamics @xcite specifies the relationship between the displacement @xmath0 that results from an applied surface traction @xmath1 and internal body force @xmath2 for two elastic states of a linear elastic body : @xmath3dv \label{equation1}\end{gathered}\ ] ] where @xmath4 is the density , @xmath5 is acceleration , the superscripts @xmath6 and @xmath7 represent the two states , and the einstein summation convention is us ed . if @xmath8 and @xmath9 then @xmath10 , and thus eq . ( [ equation1 ] ) becomes @xmath11 we shall use this theorem to determine the surface displacement ( distortion ) due to heating of an optic by , for example , partial absorption of an incident laser beam . for the first state , which we shall refer to as the `` thermal state '' and label @xmath12 , we assume that the optic is free and thus @xmath13 , and there is a non - zero body force due to the heating . since we are interested in the distortion of the end face of the optic , we choose the second state , which is often referred to as the `` auxiliary state '' and we shall label @xmath14 , to have a traction @xmath15 applied to the end face of the optic and assume @xmath16 . thus , eq . ( [ equation2 ] ) becomes @xmath17 where @xmath18 is the internal strain produced by the traction @xmath19 , and @xmath20 is the internal stress associated with the body force : @xmath21 . consider now applying time - harmonic tractions with amplitude @xmath22 @xmath23 . it is convenient to choose @xmath24 to be orthonormal , so that @xmath25 . then , expressing the surface displacement amplitude as : @xmath26 transforms the left - hand term of eq . ( [ equation3 ] ) to @xmath27 therefore @xmath28 that is , if the amplitude of the elastic response of the optic , @xmath29 , to each of the tractions @xmath30 is known then the amplitude of the distortion of the end face of the optic , @xmath31 , due to any thermal stress distribution can be calculated using eqs . ( [ equation4 ] ) and ( [ equation6 ] ) . we shall use this approach to calculate the surface distortion due to non - uniform heating of a homogeneous isotropic body for which @xmath32 where @xmath33 is young s modulus , @xmath34 is the coefficient of thermal expansion , @xmath35 is poisson s ratio , @xmath36 , and @xmath37 is the ambient temperature . . [ equation6 ] thus becomes @xmath38 to determine the distortion of the end - face using reciprocity , one must first characterize the response of the elastic system , @xmath18 , to a set of orthonormal basis tractions @xmath39}:n = 1 , .... , n$ ] using an elastostatic fem [ 11 ] . zernike functions would be a tempting choice given our cylindrical geometry , particularly as they are orthogonal to a uniform traction and thus applying the auxiliary tractions should not apply net forces to the system . however , as shown in section iv , they are not well suited to describing the surface distortion . the orthonormal basis tractions we shall use apply a non - zero ( instantaneous ) force to the optic , leading to ill - conditioning of the fem at very low frequencies . we thus used a traction frequency of @xmath40 hz as the response is independent of frequency for frequencies well below the first resonance - see [ 12 ] for example . in all of our numerical tests , we assume a cylindrical fused silica optic with height @xmath41 200 mm , radius @xmath42 170 mm , @xmath43 mpa , @xmath44 and @xmath45 k^-1^ . a radial cross section of the optic and the meshing used for the fem is shown in fig . [ figure1 ] . nodes , and is finest on the heated top surface of the test mass . ] we assume heating of the top face by 1 w of power absorbed with a gaussian - distributed flux : @xmath46\ ] ] where the beam radius @xmath47 mm , and radiative cooling of all surfaces of the optic to surroundings at 293 k. a thermal fem @xcite is used to calculate the temperature distribution , @xmath48 , resulting from the heating . the displacement amplitude for each basis function , @xmath49 , and the total displacement , @xmath50 , are then calculated using eq . [ equation4 ] and eq . [ equation8 ] . choosing a set of orthonormal functions @xmath30 that can describe the surface distortion without requiring a large number of functions , which would necessarily include high spatial frequencies , is crucial as it reduces both the number of auxiliary tractions that must be evaluated and the requirement for using a fine mesh in the fem . thus , we describe the choice of basis functions for on - axis and off - axis heating of the optic . zernike polynomials ( see appendix a ) are often used to describe cylindrically symmetric optical aberrations , as they are orthogonal over a circular disc and can be normalized . however , as shown in figure [ figure2 ] , these polynomials are not well suited to describing the distortion . , the sum of the first six zernike components @xmath52 , and the sum of the first six orthonormalized lg components @xmath53 on - axis surface distortion due to the heating can also be described using laguerre - gauss ( lg ) functions : @xmath54 % \label{equationlg}\ ] ] where @xmath55 are laguerre polynomials of order @xmath56 ( see appendix a ) , @xmath57 is the radial coordinate and @xmath58 is a free parameter . these functions are orthogonal only over the infinite plane however . symmetric orthogonalization @xcite is therefore used , as outlined in appendix b , to construct linear combinations , @xmath59 , of lg functions that are orthonormal over the end face for a given @xmath58 . in this type of orthogonalization , the difference between the new and original functions is minimized in the least - squares sense @xcite . the optimum value of @xmath58 was chosen as described in appendix c , giving @xmath60 . the six lowest - order orthonomalized - lg functions are defined in appendix d. a comparison of @xmath61 and the sum of these components in fig . [ figure2 ] shows that the lg basis is much superior to the zernike basis . the distortion due to off - axis heating can be described using the sets of functions listed below : \(a ) hermite - gauss ( hg ) functions : @xmath62 h_n \big ( \frac{\sqrt{2 } y}{r_{0y}}\big ) \exp \big[\frac{-y^2}{r_{0y}^2}\big ] % \label{equationhg}\ ] ] where @xmath63 are the ( `` physicists '' ) hermite polynomials of order @xmath64 ( see appendix a ) . these functions are orthogonal over the interval @xmath65 . we choose @xmath66 as the heat flux has a circular cross section and we shall use @xmath67 , and thus @xmath68 \(b ) generalized lg functions : @xmath69 \begin{cases } 1 \\ \sin{l \phi } \\ \cos{l \phi } \end{cases } % \label{lglp}\ ] ] where @xmath70 is the azimuthal angle , and @xmath71 for @xmath72 . we restricted the azimuthal dependence to @xmath73 due to the symmetry of the expected distortion . orthonormalized hg and generalized - lg functions were constructed , and an optimized value of @xmath74 was selected as discussed above . hg functions up to @xmath75 ( 136 functions in total ) were initially used to describe the distortion due to a heating beam that was displaced from the center of the optic according to @xmath76(0,10 mm ) , ( 10 mm , 0 ) and ( 8.7 mm , 5 mm ) . in each case , the distortion was dominated by the same 17 components , the functions for which are plotted in appendix e. a comparison of @xmath61 and the sum of the dominant 19 components is shown in fig . [ figure4 ] . and the sum of the 17 dominant orthonormalized - hg components @xmath77 for a heat flux offset of 10 mm ] orthonormalized generalized - lg functions up to @xmath78 ( 16 functions in total ) were also generated and used to describe the distortion due to a heat flux displaced from the center of the optic by 10 mm , but they yielded slightly poorer agreement with @xmath61 . in addition , since the lower order orthonormalized - hg functions appear similar to the tem@xmath79 and tem@xmath80 eigenmodes observed in optical cavities , we chose to use that basis . we now show how to use the orthonormal bases described above with reciprocity to determine the surface distortion . in each case , the equilibrium @xmath81 values were calculated for the basis tractions and then combined with the temperature distribution @xmath48 from the thermal fem to yield the amplitudes @xmath49 . while zernike polynomials are not appropriate for describing the surface distortion in the example presented here , they can be used for a reciprocity - based calculation . table i shows a comparison of the reciprocity zernike amplitudes with those calculated by decomposing the distortion predicted by the thermoelastostatic fem . .zernike amplitudes calculated using reciprocity , @xmath49 , and thermoelastostatic fem , @xmath82 , for the axisymmetric gaussian heat flux . [ cols="^,^,^",options="header " , ] the hermite gauss functions up to order @xmath83 are orthogonalized using symmetric orthogonaliazation . of these 136 modes , the 17 modes that make the largest contribution to describing the deformed surface were selected . these modes are shown in fig . [ figurec3 ]
thermoelastic distortion resulting from optical absorption by transmissive and reflective optics can cause unacceptable changes in optical systems that employ high power beams . in `` advanced''-generation laser - interferometric gravitational wave detectors , for example , optical absorption is expected to result in wavefront distortions that would compromise the sensitivity of the detector ; thus necessitating the use of adaptive thermal compensation . unfortunately , these systems have long thermal time constants and so predictive feed - forward control systems could be required - but the finite - element analysis is computationally expensive . we describe here the use of the betti - maxwell elastodynamic reciprocity theorem to calculate the response of linear elastic bodies ( optics ) to heating that has arbitrary spatial distribution . we demonstrate using a simple example , that it can yield accurate results in computational times that are significantly less than those required for finite - element analyses .
let us consider a homogeneous poisson point process @xmath1 ( ppp ) on @xmath2 with intensity @xmath3 . the plane @xmath2 is equipped with its canonical orthonormal basis @xmath4 where @xmath5 denotes the origin @xmath6 . in the sequel , the @xmath7 and @xmath8 coordinates of any given point of @xmath2 are respectively called its abscissa and its ordinate , and often denoted by @xmath9 . + from the ppp @xmath1 , baccelli and bordenave @xcite defined the _ directed spanning forest _ ( dsf ) with direction @xmath10 as a random graph whose vertex set is @xmath1 and whose edge set satisfies : the ancestor of a point @xmath11 is the nearest point of @xmath1 having a strictly larger abscissa . in their paper , the dsf appears as an essential tool for the asymptotic analysis of the radial spanning tree ( rst ) . indeed , the dsf can be seen as the limit of the rst far away from its root . + each vertex @xmath12 of the dsf almost surely ( a.s . ) has a unique ancestor ( but may have several children ) . so , the dsf can have no loop . this is a forest , a union of one or more disjoint trees . the most natural question one might ask about the dsf is whether the dsf is a tree . the answer is `` yes '' with probability @xmath3 . [ th1 ] the dsf constructed on the homogeneous ppp @xmath1 is a.s . a tree . let us remark that the isotropy and the scale - invariance of the ppp @xmath1 imply that theorem [ th1 ] still holds when the direction @xmath7 is replaced with any given @xmath13 and for any given value of the intensity of @xmath1 . + theorem [ th1 ] means that a.s . the paths in the dsf , with direction @xmath7 and coming from any two points @xmath14 , eventually coalesce . in other words , any two points @xmath14 have a common ancestor somewhere in the dsf . this can be seen on simulations of figure [ fig1 ] . [ cols="^,^ " , ] in the context of first and last passage percolation models on @xmath15 , licea and newman @xcite and ferrari and pimentel @xcite have stated that infinite paths with the same direction have to coalesce . for technical reasons , the coalescence is harder to prove in continuum models , typically when the lattice @xmath15 is replaced by a ppp on @xmath16 . this has been done by howard and newman @xcite for ( continuum ) first passage percolation models . in the same spirit , alexander @xcite studies the number of topological ends ( i.e. infinite self - avoiding paths from any fixed vertex ) of trees contained in some minimal spanning forests . + the dsf is radically different from the graphs described in the papers mentioned previously . indeed , the construction of paths in the dsf only requires local information whereas that of first and last passage paths need to know the whole ppp . + similar directed forests built on local knowledge of the set of vertices have been studied . consider the two dimensional lattice @xmath17 where each vertex is open or closed according to a site percolation process . gangopadhyay @xcite connect each open vertex @xmath9 to the closest open vertex @xmath18 such that @xmath19 ( with an additional rule to ensure uniqueness of the ancestor @xmath18 ) . athreya @xcite choose the closest open vertex @xmath18 in the @xmath20 lattice cone generated at @xmath9 and with direction @xmath7 . ferrari @xcite connect each point @xmath9 of a ppp on @xmath2 to the first point @xmath18 of the ppp whose coordinates satisfy @xmath21 and @xmath22 . in these three works , it is established that the obtained graph is a.s . a tree . + however , these three models offer a great advantage ; the choice of the ancestor of the vertex @xmath9 does not depend on what happens before abscissa @xmath23 . this crucial remark allows to introduce easily markov processes ( indexed by the abscissa ) , the use of martingale convergence theorems and of lyapunov functions . this is no longer the case in the dsf considered here . indeed , given a descendant @xmath24 of @xmath9 and its ancestor @xmath25 , the ball @xmath26 may overlap the half - plane @xmath27 and the ancestor of @xmath9 can not be in the resulting intersection . see figure [ fig : notmarkov ] for an illustration of this phenomenon . because of this , bonichon and marckert @xcite , who consider navigation on random set of points in the plane , have to restrict to cones of width at most @xmath28 for instance and their techniques do not apply to our case with @xmath29 . + notice also our discrete model bears similarities with the brownian web ( @xcite ) , for which it is known that the resulting graph is a tree with no bi - infinite path . see figure [ fig1 ] ( b ) . , @xmath25 and @xmath9 are points of the ppp @xmath1 . in the dsf with direction @xmath7 , @xmath24 and @xmath25 are some descendants of @xmath9 and @xmath25 is the ancestor of @xmath24 . the ancestor of @xmath9 can not belong to the hatched region._,width=79,height=188 ] our strategy to prove theorem [ th1 ] is inspired from the percolation literature . it is based on a comparaison argument between surface and perimeter due to burton and keane @xcite , and used in @xcite . the key properties of the point process @xmath1 that are needed are its invariance by a group of translations and the independence between the restrictions of @xmath1 to disjoint sets . this independence is to be used carefully though : for instance , modifying the ppp locally in a bounded region may have huge consequences on the paths of the dsf started in another region . + then , theorem [ th1 ] is extended to random environment . more precisely , if we delete all the points of the ppp @xmath1 which belong to an auxiliary ( and independent ) boolean model , then the dsf constructed on the remaining ppp is still a tree ( theorem [ th1boolean ] ) . + finally , theorem [ thbiinf ] states that , in the dsf with direction @xmath7 and with probability one , the paths with direction @xmath30 coming from any vertex @xmath12 are all finite . combining this with theorem [ th1 ] , this means that the dsf has a.s . one topological end . + in section [ proofpart1 ] , a key result ( lemma [ corol2 ] ) is stated from which stems the comparison argument between surface and perimeter that allows to prove theorem [ th1 ] . section [ section : p(f)>0 ] is devoted to the proof of lemma [ corol2 ] . in section [ sectiontrous ] , we show that theorem [ th1 ] still holds when the set on which @xmath1 is defined becomes a random subset of @xmath16 . finally , in section [ sectionbiinfini ] , we use the fact that the dsf is a tree to prove , as an illustration , that there is no bi - infinite paths in the dsf . for @xmath11 , let us denote by @xmath31 the path of the dsf started at @xmath12 and with direction @xmath10 . it is composed of the ancestors of @xmath12 , with edges between consecutive ancestors . this path will be identified with the subset of @xmath2 corresponding to the union of segments @xmath32 $ ] in @xmath2 where @xmath33 and @xmath34 are two consecutive ancestors of @xmath12 . by construction , this path is infinite . + let @xmath14 . either the paths @xmath31 and @xmath35 coalesce in a point @xmath36 which will be the first common ancestor to @xmath12 and @xmath24 , and coincide beyond this point . or they do not cross and are disjoint subsets of @xmath2 . this trivial remark will be widely used in section [ section : p(f)>0 ] . + let us denote by @xmath37 the number of _ disjoint _ infinite paths of the dsf . theorem [ th1 ] is obviously equivalent to @xmath38 . now , we assume that @xmath39 and our purpose is to obtain a contradiction . we first state the key result ( lemma [ corol2 ] ) which will be proved in section [ section : p(f)>0 ] and then explain how it leads to a contradiction . + before starting the proof , let us remark that the ergodicity of the ppp @xmath1 implies that the variable @xmath40 is constant almost surely . so without loss of generality , we could assume @xmath41 instead of @xmath42 . see @xcite , chapter 2.1 for details on ergodicity applied to stationary point proceses . + let @xmath43 be the set of positive integers . for any @xmath44 , let us denote by @xmath45 the cell @xmath46 . let @xmath47 be the following event : there exists a path @xmath31 in the dsf with @xmath48 that does not meet any other path @xmath35 for all @xmath49 : @xmath50 [ corol2 ] if @xmath42 then there exist some positive integers @xmath51 such that @xmath52 . let @xmath53 . let us consider the lattice @xmath54 consisting of the @xmath55 points @xmath56 where @xmath57 , and the union @xmath58 of nonoverlapping translated cells @xmath59 indexed by @xmath60 . we also define @xmath61 as the translated event of @xmath62 to the cell @xmath63 . + thanks to the translation invariance in distribution of the ppp @xmath1 , all the @xmath61 s , for @xmath64 , have the same probability . furthermore , if both events @xmath61 and @xmath65 occur with @xmath66 then there are in the dsf two disjoint infinite paths started respectively in @xmath63 and @xmath67 . hence , the number of events @xmath61 s occurring simultaneously is smaller than the number @xmath68 of edges of the dsf going out of the rectangle @xmath58 . it follows that for any @xmath69 : @xmath70 now , using the integers @xmath51 given by lemma [ corol2 ] , the probability @xmath71 is positive . we deduce that the expectation @xmath72 $ ] grows at least as @xmath73 . the contradiction comes from the next result ( lemma [ lemme : eta ] ) in which it is proved that @xmath72 $ ] is at most of order @xmath74 . this results from the fact that the expected number of edges crossing the boundary of @xmath58 should be of an order close to the perimeter of this rectangle . this concludes the proof of theorem [ th1 ] . [ lemme : eta ] for all @xmath44 there exists a constant @xmath75 ( depending on @xmath51 ) such that for all @xmath53 , @xmath76 \leq c l^{3/2 } ~.\ ] ] let us write the random variable @xmath68 as the sum @xmath77 where @xmath78 and @xmath79 respectively denote the number of edges exiting the rectangle @xmath58 and with lengths shorter and longer than @xmath80 . it suffices to prove that their expectations are of order @xmath74 . + each of the @xmath78 edges exiting @xmath58 and shorter than @xmath80 has an extremity in a strip @xmath81 of width @xmath80 all around the rectangle @xmath58 . this means @xmath78 is smaller than the number of points of the ppp @xmath1 in @xmath81 . its expectation is upper bounded by the area of the strip @xmath81 which is of order @xmath74 . + the number @xmath79 of edges exiting the rectangle @xmath58 and longer than @xmath80 is necessarily smaller than the number @xmath82 of points of @xmath83 for which the distance to their ancestor is larger than @xmath80 . to each of these points @xmath12 is associated an open half disc centered at @xmath12 and with radius @xmath80 in which there is no point of @xmath1 . otherwise , the ancestor of @xmath12 would be at a distance smaller than @xmath80 . to sum up , the points counted by @xmath82 all belong to the rectangle @xmath58 and are at distance from each other larger than @xmath80 . their number can not exceed the order @xmath84 . so do @xmath85 $ ] and @xmath86 $ ] . this proves the announced result . this section is devoted to the proof of lemma [ corol2 ] , to prove that the event @xmath47 defined in ( [ deff_mm ] ) occurs with positive probability under the hypothesis @xmath42 . the proof can be divided into two steps . first , we prove ( lemma [ 3paths ] ) that there exist positive integers @xmath51 such that with positive probability three disjoint infinite paths come from the cell @xmath45 . hence , the intermediate one , say @xmath87 , is trapped between the two other paths on the half - plane @xmath88 . then , a local modification of this event prevents any path @xmath35 with @xmath49 to touch @xmath87 . roughly speaking , we build a shield protecting @xmath87 from other paths @xmath35 s . see figure [ fig : shield ] . + to show that with positive probability three disjoint infinite paths come from a cell @xmath45 , we start with two disjoint paths ( lemma [ 2paths ] ) . for any positive integers @xmath51 , let us denote by @xmath89 the east side of the cell @xmath45 @xmath90 let @xmath91 and @xmath92 such that @xmath93 . we define the event @xmath94where @xmath95 denotes the boundary of @xmath45 . the event @xmath96 says that there exist two disjoint paths @xmath31 and @xmath35 coming from points @xmath97 and exiting @xmath45 by its east side @xmath89 . for technical reasons in the proof , we will require the parameter @xmath98 so that the two disjoint paths exit @xmath45 by crossing the rectangle delimited by the two segments @xmath99 and @xmath89 . [ 2paths]if @xmath42 then , for all @xmath91 , there exist @xmath44 with @xmath93 such that @xmath100 . the inequalities @xmath101 imply that at least one of the terms of the above sum is positive . let @xmath102 be the corresponding deterministic indices . let @xmath103 be the sequence of successive ancestors of @xmath11 in the dsf with respect to direction @xmath10 . theorem 4.6 of @xcite puts forward the existence of an increasing sequence of finite stopping times @xmath104 such that @xmath105 and the vectors @xmath106 are i.i.d . consequently , @xmath107 a.s . eventually goes out of the half - plane @xmath108 . so , @xmath109thus , there exists an integer @xmath110 large enough so that the probability in the right hand side of is positive . replacing @xmath111 by @xmath112 and @xmath113 by @xmath114 provides the announced result . we are now able to state a result similar to lemma [ 2paths ] , but with three paths instead of two . let us introduce , for @xmath91 and @xmath44 such that @xmath93 , the event @xmath115 defined as follows : there exist three disjoint paths @xmath31 , @xmath35 and @xmath116 coming from points @xmath117 and exiting the cell @xmath45 by its east side @xmath89 . [ 3paths ] if @xmath42 then , for all @xmath91 , there exist @xmath44 with @xmath93 such that @xmath118 . let @xmath31 , @xmath35 and @xmath116 be the disjoint paths given by @xmath119 . the parameter @xmath98 ensures that one of them , say @xmath31 , is trapped between @xmath35 and @xmath116 on @xmath120 and not only on @xmath121 . this precaution is needed to get the inclusion ( [ inclusionfinale ] ) in the sequel . remark also , @xmath122 will be sufficient to prove theorem [ th1 ] . lemma [ 3paths ] ( with any positive @xmath98 ) will be used to prove theorem [ th1boolean ] , in the next section . let @xmath51 be the positive integers given by lemma [ 2paths ] . first , for any integer @xmath123 , let us consider the translated event @xmath124 of @xmath125 to the cell @xmath126 , there exist two disjoint paths @xmath31 and @xmath35 coming from points @xmath127 and exiting @xmath126 by its east side . by stationarity , all these events have the same probability . if we assume that for all @xmath128 , the probabilities @xmath129 are null then the inequality @xmath130 leads to a contradiction as @xmath131 tends to infinity , since @xmath132 ( lemma [ 2paths ] ) . consequently , @xmath133 for some @xmath128 . + on this event , there exist four infinite paths started in the cells @xmath134 and @xmath135 , leaving the cells @xmath126 and @xmath136 through their east sides . we denote them by @xmath137 according to the ordinate at which they intersect the axis @xmath138 . see figure [ fig:3paths ] . now , among these four paths , at least three are disjoint . indeed , the path @xmath139 can not touch @xmath140 , otherwise by planarity , it will necessarily touch @xmath141 which is forbidden on the event @xmath142 . as a consequence , @xmath139 can not touch @xmath143 . + it suffices to replace @xmath111 with @xmath144 to conclude . _ here is a representation of the fourth paths @xmath137 corresponding to the event @xmath145 . on the axis @xmath146 , the two rectangles are the cells @xmath126 and @xmath136 . while the paths @xmath141 and @xmath140 coalesce , it remains three paths among @xmath137 which are disjoint._,width=377,height=302 ] the condition that the paths leave the cell @xmath45 through its east side is essential to obtain a contradiction in the previous proof . indeed , one could imagine that the path @xmath141 leaves the cell @xmath126 through its north side and goes into @xmath136 ( with @xmath147 ) through its south side in order to slip between @xmath140 and @xmath143 . in this case , there is nothing to prevent the paths @xmath139 and @xmath140 , respectively @xmath141 and @xmath143 , from coalescing . + it is time to prove lemma [ corol2 ] . the integers @xmath51 given by lemma [ 3paths ] and for which @xmath148 will provide @xmath149 . our idea is to modify the event @xmath150 by building a shield that will protect the cell @xmath45 and in particular the three disjoint paths given by @xmath151 . + for that purpose and for the rest of this section , we denote by @xmath152 to specify which point process satisfies the event @xmath150 . + let @xmath110 be a real number larger than @xmath153 . let us denote by @xmath154 the set @xmath155 now , let us put small balls in the set @xmath154 throughout the circle centered at the origin and with radius @xmath110 . let @xmath156 be a small positive real number and @xmath157 . we define a sequence @xmath158 of points of @xmath2 such that for any integer @xmath159 , @xmath160 has a positive ordinate and satisfies @xmath161 and @xmath162 . let @xmath163 be the smallest index @xmath164 such that @xmath165 belongs to the set @xmath166 . actually : @xmath167 let us add a last point @xmath168 with positive ordinate , abscissa equal to @xmath169 and such that @xmath170 . by symmetry with respect to the axis @xmath171 , we define the sequence @xmath172 . the balls centered at the @xmath173 s , @xmath174 , with radius @xmath156 are all included in @xmath154 and form a shield protecting the cell @xmath45 . see figure [ fig : shield ] . + _ here are the set @xmath175 and the directed spanning forest constructed on the ppp @xmath1 satisfying the event @xmath176 . the disjoint infinite paths @xmath177 of the event @xmath119 are represented in bold . the points @xmath178 are in gray . the hatched area is the strip of width @xmath98 between @xmath99 and @xmath89 that the paths @xmath179 and @xmath116 have to cross._,width=491,height=415 ] let @xmath180 be the event such that each ball @xmath181 , for @xmath174 , contains exactly one point of the ppp @xmath1 , and the rest of @xmath154 is empty . this event occurs with positive probability for any @xmath182 . in the sequel , we will denote by @xmath183 the point contained in the ball @xmath181 , @xmath174 . + let @xmath184 and @xmath185 be two independent ppp with respective intensities 1 on @xmath175 and on @xmath186 , then their superposition @xmath187 is a ppp of intensity 1 on the whole plane . hence , @xmath188 the first equality in comes from the fact that the realization of @xmath189 depends only on points in the set @xmath175 , and the second equality is due to the fact that @xmath1 and @xmath187 have the same distribution . + since deleting the points of @xmath184 does not affect the existence of the three paths @xmath31 , @xmath35 and @xmath116 of @xmath190 when this event is realized , we have @xmath191 and @xmath192 moreover , if @xmath193 is satisfied , so is @xmath194 provided the points of @xmath195 do not modify the dsf in the cell @xmath45 . this is false in general . this becomes true as soon as @xmath196 is satisfied with @xmath197 and @xmath198 , thanks to the following result . [ lengthedge ] let @xmath51 be positive integers and assume that at least two edges in the dsf start in the cell @xmath45 and exit it through its east side . then , any edge whose west vertex belongs to the cell @xmath45 has a length smaller than @xmath199 . lemma [ lengthedge ] will be proved at the end of this section . to sum up , for @xmath110 and @xmath156 as above ; @xmath200 combining with ( [ p1>0 ] ) , ( [ p2>0 ] ) and ( [ intersection ] ) it follows : @xmath201 to conclude the proof , it remains to prove the inclusion @xmath202 let @xmath203 be the three disjoint infinite paths given by @xmath150 . let us denote by @xmath31 the path which intersects the axis @xmath138 at the intermediate ordinate . on the one hand , we choose @xmath204 small enough so that the abscissa of @xmath205 , @xmath206 , is smaller than that of @xmath207 . hence , @xmath205 will prefer to have @xmath207 as ancestor rather than a point of @xmath208 . similarly @xmath209 will prefer to have @xmath210 as ancestor . on the other hand , since the abscissa of @xmath211 is larger than @xmath114 ( with @xmath212 ) , @xmath211 can not have a point of @xmath31 as ancestor . otherwise , the path of @xmath211 would cross @xmath35 or @xmath116 which is impossible . similarly , @xmath213 can not have a point of @xmath31 as ancestor . henceforth , assume the event @xmath176 satisfied . the previous construction prevents any path @xmath214 in the dsf with @xmath215 from touching @xmath31 on @xmath216 . it can neither coalesce with @xmath31 on @xmath217 since on the half - plane @xmath218 , @xmath31 is trapped between @xmath35 and @xmath116 . this leads to ( [ inclusionfinale ] ) and concludes the proof of lemma [ corol2 ] . + the section ends with the proof of lemma [ lengthedge ] . first remark that an edge whose both vertices belong to the cell @xmath45 has necessarily a length smaller than @xmath219 . now , let us focus on an edge @xmath220 in the dsf such that @xmath48 and @xmath221 . by hypothesis , there exists another edge @xmath222 such that @xmath223 and @xmath224 . if the abscissa of @xmath12 is smaller than that of @xmath33 then @xmath33 could be the ancestor of @xmath12 . so , @xmath225 otherwise , the same argument leads to @xmath226 . moreover , @xmath25 could be an ancestor of @xmath12 . so , @xmath227 this concludes the proof . let us now consider a boolean model @xmath228 where the germs are distributed following a ppp @xmath229 with intensity @xmath230 , independent of the ppp @xmath1 , and where the grains are balls with fixed radius @xmath231 : @xmath232 ( see @xcite for an exposition ) . we delete the points of @xmath1 that belong to @xmath228 . this amounts to considering the dsf with direction @xmath10 and whose vertices are the points of @xmath233 where @xmath234 . intuitively , the holes created by @xmath228 act like obstacles , which are difficult to cross and may prevent the coalescence of different paths . is such a dsf still a tree ? the answer is still `` yes '' . and coming from vertices with abscissa @xmath235 and their ancestors are represented in bold red lines . we see on ( a ) that the boolean holes mostly deviate the paths of the dsf , as in abscissa 50 ordinate 55 . however , when a path is trapped , as in abscissa 100 ordinate 70 , it has to cross the hole._,width=226,height=226 ] [ th1boolean ] the dsf constructed on @xmath233 is almost surely a tree . let us remark that theorem [ th1boolean ] holds whatever the value of the intensity @xmath236 of the boolean model . in particular , the dsf is a tree even if @xmath228 contains unbounded components . see @xcite for a complete reference on continuum percolation . + moreover , theorem [ th1boolean ] remains true when the radii are i.i.d . random variables , provided the support of their common distribution is bounded . we generalize the proof of theorem [ th1 ] . actually , thanks to the translation invariance of the joint process @xmath237 , most of arguments used in section [ proof ] still hold for the dsf constructed on @xmath233 . hence , we can assume that for any @xmath91 there exist @xmath44 such that the event @xmath238 has a positive probability ( i.e. lemma [ 3paths ] ) . it remains to prove that @xmath239 . + the main difference with the proof of theorem [ th1 ] lies in the construction of the shield ( fig . [ fig : shield ] ) . indeed , the boolean model @xmath228 could create holes in the shield so that the inclusion @xmath240 is no longer true . to avoid this difficulty , it suffices to require that @xmath228 does not intersect any ball @xmath181 involved in the shield . for that purpose , let us introduce the event @xmath241 defined as follows : there is no points of @xmath229 in @xmath242 ( it can be assumed without restriction that @xmath243 ) . + let @xmath244 . henceforth , on the event @xmath241 the boolean model @xmath228 can not intersect any ball @xmath181 with @xmath245 . the same shield as in the previous section can be constructed in order to protect the three disjoint paths given by @xmath238 . this leads to @xmath246 the arguments to prove that @xmath247 occurs with positive probability are the same as in section [ section : p(f)>0 ] . let us consider four independent ppp : @xmath184 and @xmath185 with intensity @xmath3 on @xmath175 and @xmath186 , and @xmath248 and @xmath249 with intensity @xmath236 on @xmath250 and @xmath251 . the existence of the three paths @xmath31 , @xmath35 and @xmath116 of @xmath252 is not affected if we change the point process @xmath1 in @xmath175 and if we delete the points of @xmath229 in @xmath250 : @xmath253 this inclusion becomes an equality whenever the events @xmath196 and @xmath254 are satisfied : @xmath255 ( as in ) . relations ( [ inclusionfinale2 ] ) and ( [ intersection2 ] ) imply : @xmath256 the proof ends with @xmath257 , @xmath258 and @xmath259 by construction of the dsf with direction @xmath7 , each path is infinite to the right with direction @xmath7 . a path @xmath87 of the dsf is said to be _ bi - infinite _ if it is also infinite to the left with direction @xmath30 . in other words , every point @xmath11 of a bi - infinite path @xmath87 is the ancestor of another point of @xmath1 ( which belongs to @xmath87 too ) . + as an illustration of theorem [ th1 ] , we use the fact that the dsf is a tree to prove : [ thbiinf ] there is a.s . no bi - infinite path in the dsf . considering the dsf as a subset of @xmath16 , we can define the number @xmath260 of intersection points of bi - infinite paths with a given vertical interval @xmath261 of @xmath2 : @xmath262 every bi - infinite path a.s . crosses any given vertical axis . hence , the announced result becomes : @xmath263 for a point @xmath12 of the dsf ( not necessarily a point of @xmath1 ) , let us denote by @xmath264 the subtree of the dsf consisting of the bi - infinite paths going through @xmath12 . + let @xmath261 and @xmath265 be two vertical intervals of @xmath2 such that the abscissa of @xmath261 is smaller than the one of @xmath265 . we define by @xmath266 the number of intersection points @xmath12 of bi - infinite paths with @xmath265 whose subtrees @xmath267 intersect @xmath261 : @xmath268 remark that the random variables @xmath260 and @xmath266 are integrable as soon as @xmath261 has finite lebesgue measure ( they are bounded by the number of edges of the dsf crossing this interval ) . + the idea is to state with stationary arguments that for every @xmath269 : @xmath270}\big)={{\mathbb e}}\big(\widetilde{r}_{\{0\}\times [ 0,l ] , \{x\}\times { { \mathbb r}}}\big).\ ] ] if holds then the coalescence of infinite paths ( theorem [ th1 ] ) forces both members of the equality to be zero . indeed , as a decreasing sequence of nonnegative integers , the limit of @xmath271 , \{x\}\times { { \mathbb r}}})_{x>0}$ ] , as @xmath272 , exists . it is smaller than @xmath3 by theorem [ th1 ] . thanks to and stationarity we get for every @xmath269 : @xmath273}\big)=l { { \mathbb e}}\big(r_{\{0\}\times [ 0,1]}\big).\end{aligned}\ ] ] letting @xmath274 in , the expectation of @xmath275}$ ] is necessarily @xmath276 and follows . + let us now prove . for @xmath277 and @xmath269 , all bi - infinite paths crossing @xmath278 $ ] also cross the vertical line @xmath279 . thus , @xmath280,\{x\}\times { { \mathbb r}}}$ ] is smaller than @xmath281}$ ] . so do their expectations : @xmath282,\{x\}\times { { \mathbb r}}}\big)\leq { { \mathbb e}}\big(r_{\{0\}\times [ 0,l]}\big ) .\ ] ] to show that ( [ etape8 ] ) is actually an equality , let us first remark that @xmath281}$ ] and @xmath283}$ ] have the same expectation . the bi - infinite paths crossing @xmath284 $ ] can be traced back to intersection points of bi - infinite paths with segments of the form @xmath285 $ ] , @xmath286 : @xmath287}= \widetilde{r}_{\{0\}\times { { \mathbb r}},\{x\}\times [ 0,l]}\leq \sum_{n\in { { \mathbb z}}}\widetilde{r}_{\{0\}\times [ nl,(n+1)l],\{x\}\times [ 0,l]}.\label{etape9}\ ] ] thanks to stationarity by vertical translation of @xmath288 for all @xmath286 , @xmath289,\{x\}\times [ 0,l]}$ ] and @xmath280,\{x\}\times [ -nl,-(n-1)l]}$ ] are identically distributed . thus , gives : @xmath290}\big)={{\mathbb e}}\big(r_{\{x\}\times [ 0,l]}\big)\leq { { \mathbb e}}\big(\sum_{n\in { { \mathbb z}}}\widetilde{r}_{\{0\}\times [ 0,l],\{x\}\times [ -nl,-(n-1)l]}\big)={{\mathbb e}}\big(\widetilde{r}_{\{0\}\times [ 0,l],\{x\}\times { { \mathbb r}}}\big).\ ] ] this provides and proves theorem [ thbiinf ] . finally , let us notice that an alternative proof exists based on the notion of bifurcating point . see the proof of theorem 2.2 of @xcite . let us describe an elementary model of coalescing random walks on @xmath291 . each vertex @xmath292 of @xmath293 is connected to @xmath294 or @xmath295 with probability @xmath296 , and independently from the others . this process provides a forest @xmath297 with direction @xmath7 spanning all @xmath293 . using classical results on simple random walks , one can prove that @xmath297 is a.s . a tree ( @xcite ) . + now , let us define a graph @xmath298 on @xmath299 as follows : @xmath300 is connected to @xmath301 ( resp . to @xmath302 ) if and only if the vertex @xmath303 is connected to @xmath304 ( resp . to @xmath305 ) in @xmath297 . first , the forests @xmath297 and @xmath298 are identically distributed , which implies that @xmath298 is also a tree . moreover , @xmath298 is the dual graph of @xmath297 : since @xmath298 is a tree , there is no bi - infinite path in @xmath297 . + in the same way , we wonder if the dsf constructed on the ppp @xmath1 admits a dual forest with the same distribution , which will allow to derive theorem [ thbiinf ] from applying theorem [ th1 ] to this dual forest . * acknowledgements : * the authors are grateful to franois baccelli for introducing the dsf model to them . the authors also thank the members of the `` groupe de travail gomtrie stochastique '' of universit lille 1 for enriching discussions .
we consider the directed spanning forest ( dsf ) constructed as follows : given a poisson point process n on the plane , the ancestor of each point is the nearest vertex of n having a strictly larger abscissa . we prove that the dsf is actually a tree . contrary to other directed forests of the literature , no markovian process can be introduced to study the paths in our dsf . our proof is based on a comparison argument between surface and perimeter from percolation theory . we then show that this result still holds when the points of n belonging to an auxiliary boolean model are removed . using these results , we prove that there is no bi - infinite paths in the dsf . @xmath0 keywords : stochastic geometry ; directed spanning forest ; percolation . + ams classification : 60d05
nonlinear optics has been proven to be very useful for the investigation of ferromagnetism at surfaces due to its enhanced sensitivity to twodimensional _ ferromagnetism _ @xcite . the magnetic effects are usually much stronger than in linear optics ( rotations up to 90@xmath0 , pronounced spin polarized quantum well state oscillations @xcite , magnetic contrasts close to 100@xmath1 ) @xcite . recently , second harmonic generation ( shg ) has been successfully applied to probe _ antiferromagnetism _ ( visualization of bulk af domains @xcite ) . the potential of shg to study the surface antiferromagnetism has been announced in ref . @xcite extensively discussed in our previous paper @xcite . the practical importance of studies in this field follows from the applications of antiferromagnetic ( af ) oxide layers in devices such as those based on tmr ( tunneling magnetoresistance ) , where a trilayer structure is commonly used . the central layer of tmr devices consists of an oxide sandwiched between a soft and a hard magnetic layer . for these technological applications it is necessary to develop a technique to study buried oxide interfaces . such a technique can be shg . one of the promising materials for the mentioned devices is nio . however , to the best of our knowledge , the understanding of its detailed spin structure is scarce - even the spin orientation on the ferromagnetically ordered ( 111 ) surfaces is not known . our recent paper @xcite presented an extensive study of the nonlinear electrical susceptibility tensor @xmath2 ( the source for shg within the electrical dipole approximation ) , mostly for monolayer structures . it has been proven there that the spin structure of an antiferromagnetic monolayer can be detected by means of shg . also the possibility of antiferromagnetic surface domain imaging has been presented for the first time . as it was mentioned in this previous work , bilayer spin structures are enough to account for the symmetry of a surface of a cubic antiferromagnet . here we present an extention of that work to the ( 110 ) and ( 111 ) bilayer structures , thus completing our group theoretical analysis of low miller - index antiferromagnetic surfaces . we follow exactly the group theoretical method described in ref . @xcite . at this point it is necessary to define the notions of `` phase '' and `` configuration '' , used henceforth to classify our results . `` phase '' describes the magnetic phase of the material , i.e. paramagnetic , ferromagnetic , or af . secondly , the word `` configuration '' is reserved for the description of the magnetic ordering of the surface . it describes the various possibilities of the spin ordering , which are different in the sense of topology . we describe af configurations , denoted by little letters a ) to l ) , as well as several ferromagnetic configurations , denoted as `` ferro1 '' , `` ferro2 '' , etc . the number of possible configurations varies depending on the surface orientation . all the analysis concerns collinear antiferromagnets , with one easy axis . the tables show the allowed tensor elements for each configuration . the tables also contain the information on the parity of the nonvanishing tensor elements : the odd ones are printed in boldface . in some situations an even tensor element ( shown in lightface ) is equal to an odd element ( shown in boldface ) , this means that this pair of tensor elements is equal in the domain which is depicted on the corresponding figure , but they are of opposite sign in the other domain . the parity of the elements has been checked in the operations @xmath3 , @xmath4 , and in the operation connecting mirror - domains to each other ( for the definition of the mirror - domain structure see ref . @xcite ) . the domain operation(s ) on which the parity depends is ( are ) , if applicable , also displayed in the tables . if two or more domain operations have the same effect , we display all of them together . to make the tables shorter and more easily readable some domain operations ( and the corresponding parity information for the tensor elements ) are not displayed , namely those that can be created by a superposition of the displayed domain operations . we also do not address the parity of tensor elements in the @xmath5 nor @xmath6 operations for ( 111 ) surfaces nor any other operation that `` splits '' tensor elements , although these operations also lead to a domain structure @xcite . as has been discussed in ref . @xcite it is possible to define a parity of the tensor elements for the @xmath6 and @xmath5 operations , however the tensor elements then undergo more complicated changes . the situations where the parity of the tensor elements is too complicated to be displayed in the tables are indicated by a hyphen in the column `` domain operation '' . for the paramagnetic phase , where no domains exist , we display the hyphen as well . for some configurations , none of the operations leads to a domain structure - in those configurations we display the information `` one domain '' . the reader is referred to ref . @xcite for the particularities of the parity check . the previously described af configurations of the ( 001 ) monolayer most commonly get split into two different configurations when a bilayer structure is considered . for the ( 110 ) bilayer it is not the case - only two of twelve af configurations get split in this way , thus one obtains 14 af configurations of the ( 110 ) bilayer . describing the results of our analysis we use the nomenclature of our previous article , i.e. the antiferromagnetic configurations are labeled by small letters . only the four configurations that result from splitting of the two configurations of the monolayer structure are labeled by small letters with subscripts that carry the information about how they have been constructed from the ( 110 ) monolayer . for configurations with subscript `` a '' the lower layer is constructed by translation of the topmost layer by vector ( 0.5a , 0.5b ) , where a and b are interatomic distances _ within _ the ( 110 ) plane along @xmath7 and @xmath8 axes , respectively . for configurations with subsript `` b '' the vector of translation is ( -0.5a , 0.5b ) . this corresponds to the way we constructed the ( 001 ) bilayers in @xcite . the configurations of the ( 110 ) monolayer structure are depicted in fig . [ f110 ] , and the way the bilayer is constructed is depicted in fig . [ cons110 ] . the tensor elements are presented in table [ t110 ] . in general , we can observe five types of response . however , the possibility to distinguish af configurations is not much improved compared to the ( 110 ) monolayer . even the possibility to detect the magnetic phase of the surface is not evident . as for the ( 001 ) surface @xcite , there is no difference in shg signal between the monolayer and bilayer for the paramagnetic and ferromagnetic phases . for most af configurations , however ( confs . a ) , b ) , c ) , e ) , f@xmath9 , f@xmath10 ) , g ) , h ) , j ) , k ) , and l ) ) such a difference is present due to a lower symmetry of the bilayer . in order to be consistent with our previous work @xcite we keep the same configuration names as in this earlier paper . that is why , for example , conf . b ) is not present here . the spin configurations of the ( 111 ) bilayer are constructed from the configurations of the ( 111 ) surface of our previous work in the way that the spin structure in the second atomic layer is the same as in the topmost layer , but shifted accordingly to form a hcp structure . taking into account the spin structure of the second layer causes all the af configurations to split , thus one obtains 10 af configurations of the ( 111 ) bilayer . the configurations are labeled by small letters ( indicating their `` parent '' configuration ) with subscript `` a '' if the mentioned shifting is along the positive @xmath7 axis , and `` b '' if the shifting is along the negative s@xmath11 axis . the configurations of the ( 111 ) monolayer are depicted in fig . [ f111 ] and the construction of the bilayer is depicted in fig . [ cons111 ] . the corresponding tensor elements are displayed in tab . the results are identical to those of our previous work @xcite , where the second layer of the ( 111 ) surface was treated as nonmagnetic . this means that the spin spin structure of the second layer does not play any role for shg , however the presence of the atoms in the second layer does . from our results follows that shg can probe maximally two atomic layers of the surface of cubic two sublattice antiferromagnets , and only one of the paramagnetic or ferromagnetic surface . for the ( 111 ) surface , the _ spin _ structure of the second layer does not have any influence on shg , i.e. it does not matter from the group - theoretical point of view if the investigated surface is a termination of a bulk antiferromagnet or a monolayer grown on a nonmagnetic substrate . however , these two situations can be very different from the band - theoretical point of view . 99 r. p. pan , h. d. wei , and y. r. shen , phys . b * 39 * , 1229 ( 1989 ) . t. a. luce , w. hbner , and k. h. bennemann , phys . lett . * 77 * , 2810 ( 1996 ) . a. kirilyuk , t. rasing , r. mgy , and p. beauvillain , phys . lett . * 77 * , 4608 ( 1996 ) . r. vollmer in _ nonlinear optics in metalls _ k. h. bennemann , oxford 1998 . a. v. petukhov , i. l. lyubchanskii , and t. rasing , phys . b , * 56 * , 2680 ( 1997 ) . u. pustogowa , t. a. luce , w. hbner , and k. h. bennemann , j. appl . phys . * 79 * , 6177 ( 1996 ) . r. vollmer , m. straub , and j. kirschner , j. soc . * 20 * 29 ( 1996 ) ; surf . * 352 - 354 * 937 ( 1996 ) . m. fiebig , d. frhlich , b. b. krichevtsov , and r. v. pisarev , phys . lett . * 73 * , 2127 ( 1994 ) . m. fiebig , d. frhlich , and h - j . thiele , phys . b * 54 * , r12681 ( 1996 ) . m. fiebig , d. frhlich , g. sluyterman , and r.v . pisarev , appl . . lett . * 66 * , 2906 ( 1995 ) . a. dhn , w. hbner , and k. h. bennemann , phys . lett . * 77 * , 3929 ( 1996 ) . m. trzeciecki , a. dhn , w. hbner , submitted to phys . rev . b. the notion of the parity operation requires that the original situation is restored after applying the operation at most twice . obviously , @xmath3 is such an operation , and for certain configurations @xmath4 satisfies this criterion as well ( where @xmath3 is just a symmetry operation ) . however , neither @xmath6 nor @xmath5 have this property - they must be applied at least 3 times to restore the original situation . concerning the tensor elements , for @xmath3 and mirror - operations a given tensor element may change its sign or not . upon the application of the @xmath4 operation it may become another tensor element ( i.e. its x indices become y and vice versa ) while its sign can change or remain unchanged . upon the @xmath6 operation , however , a single tensor element gets `` split '' into several tensor elements , and the notion of the conservation ( or not ) of the sign of the tensor element looses its sense . the tensor elements are described in the cartesian coordinate system , where quarters are the elementary entities . rotation by an angle other than @xmath12 and its multiples can not be described as interchanging the axes and a possible modification of their signs . however , it is possible to treat the parity of tensor elements in the @xmath6 and @xmath5 operation in a more complicated way . .[t110]nonvanishing elements of @xmath2 for all spin configurations of the ( 110 ) surface of a fcc lattice . we denote @xmath13 by ijk . odd elements are in bold if a parity operation exists . the configurations are depicted in fig . [ cols="^,^,<,^,<",options="header " , ]
extending our previous work we classify the nonlinear magneto - optical response at low index surfaces of fcc antiferromagnets , such as nio . antiferromagnetic bilayers are discussed here as models for the termination of bulk antiferromagnets . 2
the generation of classically distinguishable quantum states , a major endeavor of modern physics , has long been the object of extensive theoretical studies . in recent times important experimental investigation with atoms has been carried out in this field by various research groups @xcite@xcite@xcite@xcite@xcite . in this context it has been proved that the realization of the schroedinger - cat program is generally challenged by an extremely rapid decoherence process due to the stochastic interactions of any freely evolving _ _ mesoscopic system with the environment @xcite@xcite . within the framework of quantum computation _ _ the same process has also been recognized to represent a major limitation toward the coherent superposition of the qubits carrying the quantum information @xcite . in the domain of quantum optics several strategies have been proposed to overcome the problem , e.g. the back - action evasion @xcite and cavity control by optical feedback @xcite@xcite . in the present letter we present a new approach to the problem based on the amplifying / squeezing operation of the travelling wave optical parametric amplifier ( opa ) operating in a novel entangled configuration and initiated by a dynamical interaction process here introduced for the first time in the framework of the nonlinear ( nl ) parametric amplification : _ quantum - injection . _ in general we may define the quantum injection process in connection with any amplifying or scattering system as the one provided by an input field whose @xmath1 does not exist as a tempered solution @xcite . in our present case the character of _ quantum - injection _ is provided by the subpoissonian character of a single photon in the fock state n=1 in a quantum superposition of polarization , or momentum , states . sometimes we refer to this single particle state as the input _ qubit_. this photon may belong to a couple generated by spontaneous parametric down - conversion ( spdc ) e.g. in a @xmath2 entangled state of linear polarization @xmath3defined in a hilbert space of dimensionality @xmath4 . the spdc process has been adopted within recent tests of violation of bellinequalities @xcite , of quantum state - teleportation @xcite and of all processes generally belonging to the chapter of _ nonlocal entangled interferometry _ the key idea of the present work relates to the possibility of `` amplifying '' this quite interesting phenomenology to a higher dimensionality spin-@xmath0{0xbd}}% % beginexpansion { \frac12}% % endexpansion $ ] hilbert space , i.e. involving a large number of photon couples , by taking advantage of the _ unitary _ character of the transformation accounting for the amplification process . we show that this can be realized by a novel optical device , the quantum - injected _ , _ _ entangled opa _ leading to a new _ entangled _ schroedinger - cat ( s - cat ) configuration which may be _ decoherence - free _ , in the ideal case . a first account of the present work is found in refs.@xcite . this quite interesting condition , implying the linear superposition of two multi - particle , i.e. macroscopic_pure - states _ will be investigated for two different optical travelling - wave configurations : ( 1 ) the quantum - injected non - degenerate opa and ( 2 ) the quantum - injected mode - degenerate opa . both schemes will be analyzed theoretically by two different and complementary approaches . accordingly , the paper is organized as follows : first , the dynamical unitary evolution of the input qubit providing the quantum injection will be analyzed in the details for the two configurations in sections 2 and 3 , respectively . second , the general approach will be followed by an _ exact _ , closed form evaluation of the wigner function , in sections 4 again for both configurations . the details of all calculations will be given for the more complex and elaborate configuration ( 1 ) , for generality.rather surprisingly , we shall see that the formal expression of the wigner function is found the same for both configurations in spite of the somewhat different dynamics and of the different signification of the variables appearing in the resulting expression . this may emphasize the formal role taken by the common single - photon quantum injection scheme within the quantum analysis . in any case this allows an interesting unifying wigner function analysis showing , for both configurations , the relevant multi particle quantum superposition properties of the s - cat . third , a theory of the _ first - order _ optical correlation functions of the parametrically generated field , given in section 5 suggests , for both cases , a straightforward first - order interferometric method for a direct single - beam detection of the s - cat . in addition , the detailed theoretical investigation of the _ second - order _ inter - mode correlation function leads to a bell - type multi - photon quantum nonlocal behaviour of the emitted field . at last , a new multi - particle bell - inequality experiment will be considered . consider the diagram shown in figure 1 , a diagram suggested by an actual experimental investigation presently being carried out in our laboratory . two equal and equally oriented nl crystals , e.g. , beta - barium - borate ( bbo ) cut for type ii phase matching are excited by two beams derived from a common uv laser beam at a wavelength ( wl ) @xmath5 . crystal 1 is the spdc source of @xmath6_entangled _ photon couples emitted , with wl @xmath7 over the modes @xmath8 @xmath9 determined by two fixed pinholes . the spdc quantum - injector is provided by a type ii @xmath2 generator of linear polarization @xmath10-entangled photon couples . the crystal realizing the opa action is cut for type ii , noncollinear phase matching and is equal to the one realizing spdc . the detection system consists of a birefringent plate @xmath11 , a @xmath12 r(@xmath13 ) , a polarizing - beam - splitter pbs and two cooled photomultipliers . in the experiment a similar system is inserted on mode @xmath14 . we found that the entanglement phase @xmath15 of the output state of the couple can be easily tuned over the range @xmath16 by rotating by an angle @xmath17 the crystal around the excitation axis @xmath18 , @xmath19 being a linear function @xcite . in order to prevent any epr type state reduction that may affect the overall superposition process and then destroy the s - cat at the outset , the photon emitted over the output mode @xmath20 is filtered by a polarization analyzer with axis oriented at @xmath21to the horizontal ( t.h . ) before being detected by @xmath22 @xcite . an alternative solution for quantum injection , succesfully tested in the experiment , is provided by a type i nl crystal 1 feeding the opa by a single photon with @xmath23 oriented at 45@xmath24 , the other photon exciting @xmath25without any @xmath26 . in both cases , a click at @xmath22 opens a gate selecting all registered outcomes , thus providing the _ conditional _ character of the overall experiment . the photon emitted over @xmath27 provides the quantum - injection into the opa , physically consisting of the other nl crystal . the input state to our amplifier system may be expressed in terms of superposition of fock states associated with the modes @xmath28 @xmath29 and with the two @xmath30 components respectively parallel and orthogonal ( t.h . ) : @xmath31{0xbd}}% % beginexpansion { \frac12}% % endexpansion } \left| 0\right\rangle _ { 2\perp } \otimes \left| 0\right\rangle _ { 2\mid \mid } \otimes \left [ \left| 1\right\rangle _ { 1\perp } \otimes \left| 0\right\rangle _ { 1\mid \mid } + e^{i\phi } \left| 0\right\rangle _ { 1\perp } \otimes \left| 1\right\rangle _ { 1\mid \mid } \right]\ ] ] for a type ii nl crystal operating in noncollinear configuration the overall amplification process taking place over @xmath32 is contributed by two equal and independent amplifiers @xmath33 and @xmath34 inducing unitary transformations respectively on two couples of time dependent field operators : @xmath35 and @xmath36 for which , at the initial time of the interaction , @xmath37 is @xmath38 = \left [ \widehat{b}_{i}(0),\widehat{b}% _ { j}(0)^{\dagger } \right ] = \delta _ { ij}$ ] and @xmath39 $ ] @xmath40 for any @xmath41 and @xmath42 and @xmath43 . a quantum analysis of the dynamics of the system leads to a linear dependence of the field operators on the corresponding input quantities , e.g. for @xmath44 : @xmath45 = \left [ \begin{array}{ll } c & s \\ s & c \end{array } \right ] \left [ \begin{array}{l } \hat{a}_{1}(0 ) \\ \widehat{a}_{2}(0)^{\dagger } \end{array } \right]\ ] ] being : @xmath46 @xmath47 @xmath48 @xmath49 the _ amplification _ _ gain , _ @xmath50 the coupling term proportional to the product of the 2@xmath51-order nl susceptibility of the crystal and of the _ pump _ field , assumed classical and undepleted by the parametric interaction . the interaction time , t may be determined in our case by the length l of the nl crystal . the evolution operator for @xmath33 is then expressed in the form of the unitary _ squeeze operator _ : @xmath52 @xmath53 $ ] being : @xmath54 , @xmath55 . a corresponding @xmath56 for @xmath57 is given by the replacement @xmath58 . by use of the overall propagator @xmath59 and of the disentangling theorem @xcite , the output state is found : @xmath60 where:@xmath61;@xmath62 , @xmath63 @xmath64 @xmath65 , @xmath66 _ _ and @xmath67 @xmath68= @xmath69 is a thermal _ _ distribution accounting for the squeezed - vacuum _ _ noise _ _ with average photon number @xmath70 @xcite . details on the evaluation of ( 3 ) are given in appendix a. below the two states espressed in ( 3 ) as : @xmath71 , @xmath72 represent the effect of the one - photon quantum - injection_. _ since this sum is extended over the complete set of n - states the appeal to the _ macroscopic _ quantum coherence is justified_. _ the output state function , written in the form @xmath73 $ ] with @xmath74 and _ _ @xmath75 expresses the condition of quantum superposition between two _ pure , _ multi - particle states originating , through unitary opa transformations , from the input single - particle state @xmath76 , keeping in this process its original phase @xmath77 . in facts , any unitary transformation may generally transform but not cancel the relevant quantum properties of the input state , such as superposition and entanglement , even within a ( noisy ) process of particle amplification as in our case . furthermore , most important , since @xmath78 is not factorizable in terms of linear polarization @xmath79 it keeps his original @xmath80_entanglement _ character thus transferring into the multi - particle regime the striking quantum nonseparability and bell - type nonlocality properties of the microscopic _ ( i.e. _ 2-particle ) systems @xcite@xcite . the @xmath81_entanglement _ properties of the output state can be investigated experimetally either by a multi - particle bell inequality experiment , we shall consider in section 5 , or more simply and directly by the _ ad hoc _ optical configuration already succesfully investigated in our laboratory for the case of a @xmath6_entanglement _ single photon couple @xcite@xcite . for convenience , a layout of a possible experiment of this sort applied to our s - cat condition is reported in the inset of figure 1 . consider now the diagram shown in figure 2 : the two nl crystals , e.g. , bbo cut for type ii phase - matching , are again excited by two beams derived from the common uv source at wavelength @xmath82 . crystal 1 is again the spdc source of couples of @xmath6entangled photons with @xmath83 emitted over the two output modes @xmath84 @xmath85 determined by two fixed pinholes according to the phenomenology already discussed in the previous section . here again the photon emitted over @xmath27provides the quantum - injection into the opa consisting of the nl crystal 2 which is now cut for _ collinear _ operation over the two linear polarization modes @xmath86and @xmath87 , respectively parallel and orthogonal t.h .. the photon associated with the output mode @xmath14 is filtered by a polarization analyzer with axis at @xmath88 respect to the t.h . and then detected by @xmath89 . once again , the @xmath90 prevents any nonlocal , epr type state reduction on the correlated mode @xmath91and @xmath89 generates the gate pulse providing the _ conditional _ character of the experiment . we may note that the input , quantum injection single - photon state : @xmath31{0xbd}}% % beginexpansion { \frac12}% % endexpansion } \left [ \left| 1\right\rangle _ { 1\perp } \otimes \left| 0\right\rangle _ { 1\mid \mid } + ( \exp i\phi ) \left| 0\right\rangle _ { 1\perp } \otimes \left| 1\right\rangle _ { 1\mid \mid } \right]\ ] ] is again entangled over the two polarization modes belonging to the common momentum state @xmath27 . note , by comparison with the input state expressed by equation 1 , corresponding to the optical configuration 1 , that here two channels feeding vacuum into the nl crystal 2 are missing , since in the present _ collinear _ configuration the @xmath33 and @xmath34 collapse into a single @xmath92and the dynamics is simplified . as we shall see in section 5 this results in a larger signal - to - noise ratio of the output field . in order to investigate the properties of the output beam after filtering against uv , a birefringent plate @xmath93 and a fresnel - rhomb @xmath94 induce respectively a field s phase delay @xmath95 and a @xmath96 ) @xmath97 ( t.h . ) . then the two orthogonal @xmath23 components are detected by @xmath98and @xmath99 after separation by a polarizing beam splitter pbs . this realizes a @xmath6interferometer as we shall see . we may analyze the present amplification process following the lines of the theory given in section 2 . the single @xmath100 process induces a unitary transformation on the couple of time dependent field operators : @xmath101 and @xmath102 for which , at the initial time of the nl interaction , @xmath37 is @xmath103 = \left [ \widehat{b}(0),\widehat{b}% ( 0)^{\dagger } \right ] = 1 $ ] and @xmath104 $ ] @xmath40 . note that the present notation for the denomination of the field operators , is _ not consistent _ with the one adopted in section 1 . it can not lead to confusion but rather it is found convenient in view of the comparative discussion , in the next section , of the final expressions of wigner functions found for the two configurations . here again quantum analysis of the dynamics leads to a linear dependence of the field operators on the corresponding input quantities : @xmath105 = \left [ \begin{array}{ll } c & s \\ s & c \end{array } \right ] \left [ \begin{array}{l } \hat{a}(0 ) \\ \widehat{b}(0)^{\dagger } \end{array } \right]\ ] ] the evolution operator is then expressed in the form of the _ squeeze operator _ : @xmath52 @xmath106 $ ] being : @xmath107,@xmath108 . the use of the _ disentangling theorem_@xcite , leads to the output state : @xmath109 \label{psiout}\ ] ] where @xmath110 . there the two mutually orthogonal , interfering _ pure _ states are now given in the form : @xmath111 with @xmath66_. _ it may be useful to express the output function expressed by equation 6 in an entangled form involving two different output @xmath112vectors . in this connection a trivial example of `` entanglement swapping '' may be easily realized by separating the two , single momentum @xmath27 , linear polarization modes into two , single polarization , momentum modes @xmath9and @xmath113 , by a simple insertion of a polarizing beam splitter ( pbs ) right at the output of the opa , as shown in figure 2 . in that case the state at the output of pbs is given by : @xmath114\ ] ] @xmath115 note that the present entangled wavefunction is quite different from the one given by equation 3 and related to the optical configuration ( 1 ) . the entanglement character of the above function can be revealed by the same experimental methods referred to at the end of section 1 . the optical configurations presented in figures 1 and 2 are not the only possible nor the more convenient ones . consider for instance that the mode - degenerate , collinear optical configuration shown in figure 2 can be easily replaced by a less elegant , _ non - collinear _ configuration in which the opa , consisting of a type ii nl crystal , is fed by a quantum injection single - photon over two different input @xmath32-vectors , say @xmath116 , each @xmath32-vector corresponding to either one of the two mutually orthogonal linear polarizations @xmath117 the input single - photon state is then : @xmath118{0xbd}}% % beginexpansion { \frac12}% % endexpansion } \left [ \left| 1\right\rangle _ { 1\perp } \otimes \left| 0\right\rangle _ { 2\mid \mid } + ( \exp i\phi ) \left| 0\right\rangle _ { 1\perp } \otimes \left| 1\right\rangle _ { 2\mid \mid } \right ] $ ] . it is easy to recognize that , in this case , we are led to the _ same _ results given by equations 8 with the output mode labels 3 and 4 replaced by 1 and 2 . alternatively , the opa may consist of a type i crystal fed by quantum injection over the two @xmath119-vectors with _ equal _ polarizations @xmath120 . in this case the input state is simply expressed by @xmath118{0xbd}}% % beginexpansion { \frac12}% % endexpansion } \left [ \left| 1\right\rangle _ { 1}\otimes \left| 0\right\rangle _ { 2}+(\exp i\phi ) \left| 0\right\rangle _ { 1}\otimes \left| 1\right\rangle _ { 2}\right ] \ $ ] and again the output state is given by equations 8 . in addition , consider the optical configuration presented in figure 3 . it is a relevant three crystal variant of the above schemes and implies a _ double quantum injection _ into the opa _ , _ e.g. by adoption of two distinct spdc processes feeding in _ entangled - state _ the two input modes of the common entangled amplifier , @xmath32 @xmath121within a _ double - conditional_.experiment . in spite of the increased experimental complications , mainly due to the low probability of the simultaneous opa quantum - injection processes , this new configuration may present definite advantages . for instance , the _ signal to noise _ ratio and then the _ visibility _ @xmath122 of the schroedinger - cat field are far larger that the one for the mode non - degenerate opa , as we shall see in the next sections . in addition and most important , the feeding if the common entangled opa by two _ simultaneous _ quantum - injection processes with different phases @xmath123 adds new interesting quantum features to the output state of the emitted field . we postpone to an _ ad hoc _ paper an exhaustive analysis of such interesting complex optical configuration . in order to inspect at a deeper lever the above results , consider the wigner function of the output field for the more complex configuration 1 , shown in figure 1 . evaluate first the symmetrically - ordered characteristic function of the set of complex variables @xmath124 , @xmath125 : @xmath126d[\eta _ { 2}(t)]d[\xi _ { 1}(t)]d[\xi _ { 2}(t)]\left| \psi _ { 0}\right\rangle $ ] expressed in terms of the _ displacement _ operators : @xmath127\equiv $ ] @xmath128 $ ] , @xmath129 $ ] @xmath130 - @xmath131 $ ] where : @xmath132 ; @xmath133 ; @xmath134 ; @xmath135 . the wigner function , expressed in terms of the corresponding complex phase - space variables @xmath136 , @xmath137 , @xmath138 , @xmath139 is the eight - dimensional fourier transform of @xmath140 , namely : @xmath141\right\}\end{aligned}\ ] ] where @xmath142 , etc.@xmath143by a lengthy application of operator algebra and integral calculus we could evaluate analytically in closed form either @xmath140 and @xmath144 . _ _ the corresponding detailed calculations for the optical configuration ( 1 ) are reported in the appendix b and c. the final _ exact _ expression of the wigner function is : @xmath145\ ] ] where @xmath146{0xbd}}% % beginexpansion { \frac12}% % endexpansion } ( \gamma _ { a+}-i\gamma _ { a-})$ ] , @xmath147{0xbd}}% % beginexpansion { \frac12}% % endexpansion } ( \gamma _ { b+}-i\gamma _ { b-})$ ] are expressed in terms of the squeezed variables _ : _ @xmath148 ; @xmath149 ; @xmath150 ; @xmath151 . the wigner functions @xmath152 \right ) $ ] ; @xmath153 \right ) $ ] definite positive over the 4 - dimensional spaces @xmath154 and @xmath155represent the effect of squeezed - vacuum , i.e. emitted respectively by opa@xmath156 and opa@xmath157 in absence of any injection . inspection of equation 10 shows that precisely the quantum superposition character of the injected state @xmath158 determines the dynamical quantum superposition of the devices @xmath33and @xmath159 the ones that otherwise act as _ uncoupled _ and _ independent _ objects . from another perspective , since the quasi - probabilty functions @xmath160 , _ _ @xmath161 _ _ @xmath155corresponding to the two macrostates @xmath162and @xmath163 in absence of quantum superposition are defined in two totally separated and independent spaces , their respective distance in the overall phase - space of the system @xmath164 can be thought of as macroscopic , as generally required by any standard s - cat dynamics in a 2 - dimensional phase - space @xcite . the link between the spaces @xmath165 and@xmath166 is provided by the quantum superposition term in equation 10 @xmath167 $ ] . this term provides precisely the first - order quantum interference of the macrostates @xmath168and @xmath169 . in addition , and most important , equation 10 and figure 3 show the non definite positivity _ _ of @xmath144 over its definition space . this assures the overall quantum character of our multiparticle , quantum - injected amplification scheme @xcite@xcite . we may recognize that these last properties of the overall wigner function of our system do indeed coincide with the formal requirements of any schroedinger - cat apparatus , which may be outlined as follows @xcite : \(a ) the ability of the system to create a first - order interference fringe pattern is a necessary but _ not sufficient _ condition for any authentic s - cat behaviour . the following are indeed the only two _ necessary and _ _ sufficient _ conditions . \(b ) the wigner function defined in the overall _ phase - space _ of the system _ must not _ be definite - positive on his definition domain @xcite . \(c ) the two interfering macrostates , identified by two corresponding gaussian - like peaks of the wigner function must be clearly distinguishable , i.e. , the _ distance _ between the peaks must be larger than their average width . note that conditions ( b ) and ( c ) imply necessarily the system s ability to produce a first - order interference pattern while the inverse argument is not necessarily true , as said . all these formal s - cat properties are shown by the tridimensional plots of the reduced wigner functions given in figure 4 . these are drawn for @xmath170 and for different values of the injection phase @xmath77 , in correspondence with the configuration shown in figure 1 and investigated over the output mode @xmath14 by the detection system shown in the inset of the same figure . the wigner function of the output field related to the quantum - injected _ mode - degenerate opa _ considered in figure 2 and section 3 may be obtained by a similar theoretical analysis . in facts this one is simpler because the _ collinear _ optical configuration adopted within the nl interaction makes the two different amplifiers @xmath171 and @xmath172 of the previous case to collapse into one . it follows that the dynamics of the system , instead of being described in an eight dimensional phase - space , as in the previous case , is described here in a more handy _ four dimensional _ space , i.e. identified by two orthogonal @xmath173states of the field emitted over a _ single _ output mode , @xmath27 . for the sake of completeness let s outline the theory in the _ collinear _ case . assume the _ phase - space variables _ @xmath174 , the conjugated variables @xmath175and evaluate the the _ symmetrically _ ordered _ characteristic function _ : @xmath176 @xmath177 @xmath178d[\xi ( t)]\left| \psi _ { 0}\right\rangle $ ] expressed in terms of the operators @xmath179\equiv \exp [ \eta ( t)\hat{a}(0)^{\dagger } -\eta ^{\ast } ( t)\hat{a}(0)]$ ] , @xmath180\equiv \exp [ \xi ( t)\widehat{b}(0)^{\dagger } -\xi ^{\ast } ( t)\widehat{b}% ( 0)]$ ] where @xmath181 ; @xmath182 and @xmath183 , @xmath184are the field operators associated to the two orthogonal input @xmath173modes . the wigner function is the fourth dimensional fourier transform of @xmath185.@xmath143 once again , it is evaluated analytically in _ closed form _ and _ _ the result is found to reproduce exactly the one expressed by equation 10 , after a previous multiplication by @xmath186 . this unexpected result reached by the analysis of two entirely different optical configurations suggests that the actual form of @xmath144 given by equation 10 is indeed determined by the peculiar character of the single - photon _ quantum - injection _ process which is common to both configurations . of course the parameters appearing in the expression of @xmath187given by equation 10 in correspondence with the _ collinear _ case are now to be re - defined appropriately : @xmath188 , @xmath189 , @xmath190 , @xmath191 . relevant information about the quantum mechanical features of the quantum - injected opa systems at hand are also revealed by the @xmath192 and @xmath193 correlation functions of the output fields for both optical configurations ( 1 ) and ( 2 ) corresponding repectively to the mode non - degenerate and to the mode - degenerate cases @xcite . let s analyze both cases in correspondence with the photo - detection measurements of the output fields carried out by apparatuses equal to the one appearing in the inset of figure1 . such measurement devices work as follows . before detection over the output momentum mode @xmath32 @xmath125 the fields are phase - shifted by @xmath194 by a birefringent plates and filtered by @xmath23-analyzers with axes oriented at the angles : @xmath195 ( t.h . ) . each @xmath23-analyzer may consist of the combination of a fresnel - rhomb @xmath23-rotator , @xmath94 and of a polarizing beam splitter , pbs : figures 1 and 2 . the field associated with the mode @xmath32 is detected at the space - time positions @xmath196 by two _ linear _ detectors @xmath197 and @xmath198 with @xmath199 . the @xmath200 correlation - functions @xmath201 are ensemble averages of the number operators @xmath202 written in terms of the the detected output fields : @xmath203$],@xmath204=\delta _ { ij}$ ] , @xmath205{0xbd}}% % beginexpansion { \frac12}% % endexpansion } ( \cos \varphi _ { j}+\sin \varphi _ { j})\exp ( i\psi _ { j\beta } ) $ ] , @xmath206{0xbd}}% % beginexpansion { \frac12}% % endexpansion } ( \cos \varphi _ { j}-\sin \varphi _ { j})\exp ( i\psi _ { j\alpha } ) $ ] , where @xmath207,@xmath208 are phase - shifts induced by the birefringent plate on the fields @xmath209 , @xmath210 . these fields are determined by the linear transformations given by equations 2 and 5 for the two configurations . in our cases @xmath211 show the expected superposition character of the output field with respect to the @xmath23-rotation angles @xmath212 and to the @xmath213 being @xmath214 the phase affecting the field s output entangled state expressed by equations 3 and 6 . the explicit evaluation of the first order functions for zero time delay leads to the following results in correspondence with the optical configurations ( 1 ) and ( 2 ) . the @xmath215 functions@xmath216 are also found for simultaneous photo - detection processes on two equal or different * k-*modes @xmath41,@xmath42 @xmath217 . the@xmath200 correlation functions are found to be expressed by : @xmath218{0xbd}}% % beginexpansion { \frac12}% % endexpansion ( \overline{n}+1)[1+\cos ( 2\varphi _ { 1})\cos \delta \qatop{-}{1}\phi ] $ ] , @xmath219 @xmath220{0xbd}}% % beginexpansion { \frac12}% % endexpansion $ ] @xmath221 $ ] . these averages are related correspondingly to photodetection measurements carried out over the output modes 1 or 2 by the detection apparatus shown in the inset of figure 1 . we may compare these results with the corresponding averages over the input _ vacuum state , _ i.e. , in.absence of any quantum - injection process . these averages are found to be independent of @xmath222 and @xmath223 , as expected and account for the unavoidable , _ squeezed _ _ vacuum _ quantum noise affecting our active parametric method : @xmath224 . by this comparison we obtain the _ signal - to - noise - ratio _ related to the schroedinger - cat detection : @xmath225 @xmath177 @xmath226 for @xmath227 . the above result immediately suggests a @xmath192-@xmath228interferometric method for s - cat detection on a _ single _ @xmath229 beam _ , _ with _ visibility : _ @xmath230 _ = _ @xmath231 . the wigner function plotted in figure 4 refer to the output field detected by this method on the mode @xmath14 . note that the average _ difference _ between the signals obtained at the output of the detectors @xmath197 and @xmath198 placed at the output arms of pbs of the apparatus , figure 1 inset , operating on the output mode @xmath32 , expresses directly the _ fringe pattern _ related to the @xmath200 interference between the two s - cat macrostates . for instance for @xmath232 , we have : @xmath233= @xmath234 . the @xmath215 functions@xmath235 are also found : @xmath236= @xmath237\right\ } $ ] ; @xmath238 @xmath239 \right\ } $ ] ; @xmath240= @xmath241+@xmath242+@xmath243$]+ @xmath244 @xmath245 $ ] + @xmath246\cos ^{2}\delta \varphi ^{-}\right . $ ] @xmath247 @xmath248 $ ] @xmath249 where : @xmath250 , @xmath251.we may easily prove , e.g. for all @xmath252= @xmath212=@xmath253,@xmath143that our system realizes the _ maximum _ quantum mechanical violation of the cauchy - schwarz inequality which generally holds in semi - classical field theory : @xmath254^{2}\leq g\qatop{(2)}{11}(0)~g\qatop{(2)}{22}(0)$ ] being : @xmath255 @xmath256^{-1}$]@xcite . furthermore , the given expression of @xmath257 shows the effects of the _ multiparticle _ quantum nonseparability and bell - type nonlocality , contributed by the terms proportional to @xmath258 and @xmath259 . this is a most relevant manifestation of the nonlocality properties of our quantum injected , _ entangled _ parametric system @xcite@xcite . the analysis above can be repeated for the simpler dynamics of the _ collinear _ case , configuration ( 2 ) . the results are : @xmath200 correlation function : @xmath260{0xbd}}% % beginexpansion { \frac12}% % endexpansion ) [ 1+\cos ( 2\varphi ) \cos \delta \phi ] $ ] , with @xmath261 . this leads to an _ interference _ _ fringe visibility : _ @xmath262{0xbd}}% % beginexpansion { \frac12}% % endexpansion , $ ] for @xmath263 . note that in the collinear case @xmath122 is larger by a factor @xmath264 respect to the noncollinerar case , a result due to the absence in the dynamics of the input vacuum fields contributions over the mode @xmath27 , figure1 . the absence of an input vacuum field on the idler mode is a most favorable condition shared also by the three - crystal opa configuration shown in figure 3 . again the _ fringe pattern _ related to the 1@xmath265 order interference between the two s - cat macrostates is determined by the difference : @xmath266 @xmath267 the @xmath268-@xmath269 correlation functions @xmath270_{\varphi } [ % \widehat{n}(t)]_{\varphi ^{\prime } } : \left| \psi _ { 0}\right\rangle $ ] may also be measured by use of the detectors @xmath98 and @xmath271 , [email protected] expressions are given here for completeness : @xmath273 @xmath274+@xmath275 + @xmath276 + @xmath277 ; @xmath278= @xmath279 - @xmath280 . the virtual absence of any effective decoherence within the travelling - wave ( tw ) parametric process we are considering may be understood by the following argument . our schroedinger - cat system does not consist of a _ free _ excitation _ , _ a condition common to _ all _ s - cats considered thus far in the literature . it consists in fact of a _ driven _ excitation which it is strongly coupled with a _ continuously _ _ re - phasing _ environment provided by the coherent _ nonlinear _ polarization of the parametric process _ driving _ the multiphoton field in quantum superposition . similar situations of coherence persistence _ in spite _ of dissipation are often encountered in physics of the _ nonlinear _ dynamical systems , e.g. , in solid state nonlinear spectroscopy . a nice example is provided there by the nonlinear generation of surface - polaritons or plasmons in strongly light absorbing semiconductors or metals @xcite . in spite of any arbitrarily large damping of the medium , high intensity and _ strictly _ _ nondecaying , _ _ driven _ polariton waves _ _ may be _ _ nonlinearly _ _ generated over arbitrarily _ _ large distances and times while the related _ free _ waves _ , _ that originate at the boundaries from the _ _ driven _ _ ones _ , _ die out quickly according to the linear optical properties of the medium . note that in the _ linear _ regime , i.e. where the only driving force is the linear polarization , the two kinds of waves coincide and are damped at the same rate . we believe that the above interpretation is generally valid for any kind of dissipative process , e.g. damping , de - phasing and de - coherence . since this one is a most disruptive process for quantum coherence in complex systems , e.g. in the domain of _ quantum computation _ , our results would then suggest the _ nonlinear _ interaction among the information carrying particles as a most efficient solution toward a large scale implementation of the new methods @xcite . of course any single photon loss event , mainly contributed in the present tw case by stray reflections , implies an elementary decoherence process . in our laboratory experiment two equal 1 mm thick , bbo crystals are excited by 0.8 picosecond pulses at @xmath281 second - harmonic - generated by a mode - locked ti : sa laser at a 76 mhz repetition rate with an average power @xmath282 . the detection system , consisting of two linear photodetectors connected to an electronic correlator , is equal to the one shown by figure1 inset , but for the absence of the birefringent plate . the initial phase is : @xmath283 all surfaces are treated by special ar coatings resonant at the working @xmath284 with an overall transmittivity : @xmath285 . this figure implies the loss of a single photon every @xmath286 pulses with the generation of @xmath287 per pulse . this would make our s - cat experiment quite feasible . note in this connection that the _ travelling - wave _ case is quite superior to the optical parametric _ oscillator _ ( opo ) configuration where the presence of unavoidable cavity losses enhance the negative effect of all phase - disrupting processes @xcite . in summary , we have given the quantum analysis of a novel nonlinear entangled tw parametric system that shows macroscopic , decoherence free , quantum superposition features that can be easily detectable . this result is reached by a smart interplay of the fundamental paradigms of modern quantum optics , i.e. , _ quadrature squeezing _ , _ multiparticle state entanglement _ and _ quantum nonseparability _ in parametric correlations . from a foundational perspective , our method could find application within the realization of fundamental nonlocality and noncontestuality tests of quantum mechanics requiring a number of entangled particles larger than two @xcite . in addition , within the fields of _ quantum information _ and _ computation _ it may represent a new way to amplify quantum coherence and entanglement over large systems providing at the same time an elegant way to beat decoherence . of course we are dealing here with a _ noisy _ system . however the quantum - noise - reduction effect contributed efficiently by the parametric _ quadrature squeezing _ may find here a successful application @xcite . _ _ for many reasons we are inclined to believe that , if successful , the present project may open new and long reaching paths in some fundamental chapters of modern physics . we thank s.branca , m.dariano , g.di giuseppe , d.p.di vincenzo , g.ghirardi for enlightening discussions , the cee - tmr program ( contract n.erbmrxct96-066 ) and infm ( contract pra97-cat ) for funding . the application of the _ disentangling theorem _ in the context of the present work to the general input fock state @xmath288implies the use of the following transformations . @xmath106(\left| n\right\rangle _ { 1\perp } \otimes \left| m\right\rangle _ { 2\mid \mid } ) _ { in}=\exp \gamma \widehat{% a^{\dagger } } \ exp(\varsigma \lbrack \widehat{a^{\dagger } } , \widehat{a}% ] ) \exp ( -\gamma \widehat{a})(\left| n\right\rangle _ { 1\perp } \otimes \left| m\right\rangle _ { 2\mid \mid } ) _ { in}$ ] , where : @xmath289 . since @xmath290=-\{\hat{a}_{1}(t)^{\dagger } \hat{% a}_{1}(t)+\hat{a}_{2}(t)^{\dagger } \hat{a}_{2}(t)+1\}$ ] the following results are found for three relevant input states . @xmath291 @xmath106(\left| 0\right\rangle _ { 1\perp } \otimes \left| 0\right\rangle _ { 2\mid \mid } ) = exp\{-\varsigma + \gamma \widehat{a^{\dagger } } \}(\left| 0\right\rangle _ { 1\perp } \otimes \left| 0\right\rangle _ { 2\mid \mid } ) \ = \sum \qatop{\infty } { n=0}\sqrt{p_{n}}% \left| n\right\rangle _ { 1\perp } \otimes \left| n\right\rangle _ { 2\mid \mid } = $ ] @xmath292 , where @xmath293 expresses the _ thermal distribution _ of the squeezed - vacuum on the two output modes with average photon number : @xmath70 . @xmath294(\left| 1\right\rangle _ { 1\perp } \otimes \left| 0\right\rangle _ { 2\mid \mid } ) = \exp \gamma \widehat{% a^{\dagger } } \ exp(-\varsigma \lbrack \widehat{a^{\dagger } } , \widehat{a}% ] ) ( \left| 1\right\rangle _ { 1\perp } \otimes \left| 0\right\rangle _ { 2\mid \mid } ) \ = $ ] @xmath295 @xmath296 @xmath297(\left| 1\right\rangle _ { 1\perp } \otimes \left| 1\right\rangle _ { 2\mid \mid } ) = \exp ( -3\varsigma ) \sum \qatop{\infty } { n=0}\frac{\gamma ^{n}}{n!}(\hat{a}_{1}^{\dagger } \hat{a% } _ { 2}^{\dagger } ) ^{n}(\left| 1\right\rangle _ { 1\perp } \otimes \left| 1\right\rangle _ { 2\mid \mid } ) $ ] - @xmath298-@xmath299 . the other states appearing in equation 3 are obtained by the same transformations upon the substitutions : a@xmath300 @xmath301etc . the symmetrically ordered characteristic function or the optical configuration ( 1 ) is evaluated by the average : @xmath302 @xmath303d[\eta _ { 2}(t)]d[\xi _ { 1}(t)]d[\xi _ { 2}(t)]\left| \psi _ { 0}\right\rangle $ ] expressed in terms of the displacement operators : @xmath304 $ ] @xmath305 $ ] ; @xmath306\equiv \exp [ \xi _ { j}(t)\widehat{b}_{j}(0)^{\dagger } -\xi _ { j}^{\ast } ( t)\widehat{b}_{j}(0)]$ ] and of the time dependent parameters : @xmath307 ; @xmath133 ; @xmath308 ; @xmath309 where @xmath310 , @xmath311=@xmath312 represents the set of the eight phase - space conjugate complex variables relative to our dynamical problem . note that the expression of @xmath313 may be given in an equivalent , alternative form by use of the following results : @xmath314d[\eta _ { 2}(t)]=d[\eta _ { 1}]d[\eta _ { 2}]$ ] and @xmath315d[\xi _ { 2}(t)]=d[\xi _ { 1}]d[\xi _ { 2}]$ ] where : @xmath316\equiv \exp [ \eta _ { j}\hat{a}% _ { j}(t)^{\dagger } -\eta _ { j}^{\ast } \hat{a}_{j}(t)]$ ] ; @xmath317\equiv \exp [ \xi _ { j}\widehat{b}_{j}(t)^{\dagger } -\xi _ { j}^{\ast } \widehat{b}% _ { j}(t)]$ ] these last results are obtained by use of eqs.2 and of the well known theorem : @xmath318 @xmath319 @xmath177 @xmath320{0xbd}}% % beginexpansion { \frac12}% % endexpansion \left [ \widehat{a},\widehat{b}\right ] $ ] . we may then evaluate the explicit expression of @xmath176 by use of the explicit form of the input state eq.1 and of the well known relations involving _ displacement operators : _ @xmath321 , @xmath322 , @xmath323 , @xmath324{0xbd}}% % beginexpansion { \frac12}% % endexpansion \mid \alpha \mid ^{2})$ ] @xcite . by ensemble averaging over the two superposition terms appearing in eq.1 , the symmetrically ordered characteristic function is finally found : @xmath325{0xbd}}% % beginexpansion { \frac12}% % endexpansion \mid e^{i\phi } \eta _ { 1}(t)+\xi _ { 1}(t)\mid ^{2}\right\ } \exp [ -% % tcimacro{\unicode[m]{0xbd}}% % beginexpansion { \frac12}% % endexpansion \sum_{j}(\mid \eta _ { j}(t)\mid ^{2}+\mid \xi _ { j}(t)\mid ^{2})]\ ] ] in view of the explicit form of @xmath326the @xmath327-dimensional integral , equation 9 is evaluated by introducing first the change of variables : @xmath328 , @xmath329and then by writing:@xmath330 , @xmath331 , @xmath332;@xmath333where @xmath334and@xmath335 , @xmath336 . for integration purposes this transformation is completed by multiplication of the integrand by the determinant of the @xmath337jacobian matrix,@xmath338 \ $ ] with elements : @xmath339with obvious notation for the partial derivatives . it is convenient to re - define here , by a multiplication by @xmath340 , the linear transformations between the two sets of variables , i.e. : @xmath341 $ ] ; @xmath342 $ ] ; @xmath343 $ ] ; @xmath344 $ ] . the value of the jacobian determinant is then found = 16 . by the above substitutions , the argument of the exponential in the integrand of the integral , eq.9 may be cast in the form : @xmath345\ = \ \varphi _ { j}+q_{j}\sin \varphi _ { j}+\overline{p}_{j}\cos \overline{\varphi } % _ { j}+\overline{q}_{j}\sin \overline{\varphi } _ { j}]$ ] , where : @xmath346 $ ] ; @xmath347 $ ] . the couple @xmath348is obtained by entering the substitutions : @xmath349and @xmath350 in the expressions for @xmath351 . likewise , the couples @xmath352 and @xmath353are found by the substitutions : @xmath354 in the expressions for @xmath355 and @xmath356 . consider now the following expressions : @xmath357=@xmath358 , @xmath359 , where : @xmath360 $ ] , @xmath361 $ ] and @xmath362 is obtained again by the substitutions : @xmath363 . turn now to equation 9 and integrate first respect to the phases @xmath212 , @xmath364 . this step is accomplished by the use of the following results : @xmath373\ & = & \ \pi ^{-8}\iiiint d^{2}\eta _ { 1}d^{2}\eta _ { 2}d^{2}\xi _ { 1}d^{2}\xi _ { 2}f(x_{j},y_{j } ) \\ & & \ \ \ \ \ \ \ \ \ \ \ \ \ \exp \left\ { \sum_{j}[(\eta \qatop{\ast } { j}% \alpha _ { j}-\eta _ { j}\alpha \qatop{\ast } { j})+(\xi \qatop{\ast } { j}\beta _ { j}-\xi _ { j}\beta \qatop{\ast } { j})]\right\}\end{aligned}\ ] ] @xmath374let @xmath375 $ ] and consider the integral [ gr 4 , pag . 717 ] : @xmath376 . assuming @xmath377 , @xmath378 , @xmath379 this result leads immediately to : @xmath380=16\pi ^{-4}\exp [ -% % tcimacro{\unicode[m]{0xbc}}% % beginexpansion { \frac14}% % endexpansion \sum_{j}(\left| \delta _ { j}\right| ^{2}+\left| \overline{\delta } _ { j}\right| ^{2})]$ ] . @xmath381let @xmath382 $ ] and consider the integral [ gr 6.631 , pag.716 ] : @xmath383{0xbd}}% % beginexpansion { \frac12}% % endexpansion ) $ ] @xmath384{0xbd}}% % beginexpansion { \frac12}% % endexpansion ( \nu + \mu + 1)}{\bf \gamma } ( \nu + 1)]^{-1}{\bf \phi } [ % % tcimacro{\unicode[m]{0xbd}}% % beginexpansion { \frac12}% % endexpansion ( \nu + \mu + 1),\ \nu + 1;\ -\beta ^{2}/(4\alpha ) ] \ $ ] being @xmath385 a _ degenerate hypergeometric _ function . apply the kummer theorem : @xmath386 @xmath387 to the standard infinite series expansion : @xmath388 @xmath389[gr 9.210 and 9.212 ] . note that with the values of the parameters : @xmath377 , @xmath390 , @xmath378 , @xmath391 only the first two terms of the expansion are nonzero , i.e. @xmath392{0xbc}}% % beginexpansion { \frac14}% % endexpansion \left| \delta \right| ^{2})$]= @xmath393{0xbc}}% % beginexpansion { \frac14}% % endexpansion \left| \delta \right| ^{2}]\exp ( -% % tcimacro{\unicode[m]{0xbc}}% % beginexpansion { \frac14}% % endexpansion \left| \delta \right| ^{2})$ ] . this leads to the _ exact _ result : @xmath380\ = 16\pi ^{-4}[2-% % tcimacro{\unicode[m]{0xbc}}% % beginexpansion { \frac14}% % endexpansion ( \left| \delta _ { j}\right| ^{2}+\left| \overline{\delta } _ { j}\right| ^{2})\exp [ -% % tcimacro{\unicode[m]{0xbc}}% % beginexpansion { \frac14}% % endexpansion \sum_{j}(\left| \delta _ { j}\right| ^{2}+\left| \overline{\delta } _ { j}\right| ^{2})]$ ] @xmath394let @xmath395 $ ] and consider the integral gr 6.631 just given at paragraph @xmath396 . with the new set of parameters : @xmath397 , @xmath398 , @xmath378 , @xmath399 , and by use of the quoted kummer theorem within the series expansion for @xmath400{0xbc}}% % beginexpansion { \frac14}% % endexpansion \left| \delta \right| ^{2})$ ] , the _ exact _ value of the integral is found : @xmath401{0xbc}}% % beginexpansion { \frac14}% % endexpansion \left| \delta \right| \exp ( -% % tcimacro{\unicode[m]{0xbc}}% % beginexpansion { \frac14}% % endexpansion \left| \delta \right| ) $ ] . this leads to the result : @xmath402{0xbc}}% % beginexpansion { \frac14}% % endexpansion \left| \delta _ { j}\right| ) $ ] . a further calculation of an identical integral involving the variables @xmath371 , @xmath403 , @xmath404leads to the result : @xmath405\ = \ -(16\pi ^{-4})\delta _ { j}\overline{\delta } \qatop{\ast } { j}\exp [ -% % tcimacro{\unicode[m]{0xbc}}% % beginexpansion { \frac14}% % endexpansion \sum_{j}(\left| \delta _ { j}\right| ^{2}+\left| \overline{\delta } _ { j}\right| ^{2})]$ ] . at last we insert the explicit expression of @xmath406 found in appendix b within the integral expressed by equation 9 and use the above results @xmath407 , @xmath408 , @xmath409 . this leads to the final expression of @xmath144 given by equation 10 . j. franson , phys.rev.lett.62 , 2205 ( 1989 ) ; p.kwiat , a.steinberg and r.chiao , phys.rev.a , 45 , 7729 ( 1992 ) ; t. larchuk , r. campos , j. rarity , p. tapster , e. jakeman , b. saleh , and m. teich , phys.rev.lett . 70 , 1603 ( 1993 ) ; t.b . pittman , d. strekalov , a. sergienko , m. rubin , d. klyshko and y. shih , phys.rev.a 52 , r3429 ( 1995 ) ; k. mattle , h. weinfurter , p. kwiat and a. zeilinger , phys . 76 , 4656 ( 1996).d . boschi , f. de martini and g. di giuseppe in _ fundamental problems in quantum theory , _ ed . by shih and rubin fortschritte der physik , vol . 46 , 643 , 1998 . d.boschi , f.de martini and g.di giuseppe in _ quantum interferometry , _ f.de martini , g.denardo and y.shih eds.(vch , weinheim , 1996 ) . s.branca , d.boschi , f.de martini and g. di giuseppe , subm . to physics letters a. we have found experimentally that the _ entanglement _ _ phase _ @xmath214 of the spdc generated entangled 2-photon state can be tuned by rotating the nl crystal by an angle @xmath410 around the excitation axis @xmath411 , being @xmath412 @xmath410 . in bbo is found : @xmath413 at @xmath414 . f. de martini , phys . 81 ( 1998 ) and : phys.lett . a 250 , 15(1998 ) . a theory of the present process in a less convenient _ parametric oscillator _ structure is given in : f. de martini , m. fortunato , p. tombesi and d.vitali , phys.rev.a , may 6 , 1999 . a. einstein , b. podolsky , n. rosen , phys.rev . * 47,*777 ( 1935 ) ; j.s . bell , physics , 1 , 195 ( 1964 ) ; another solution for quantum injection consists of a type i nl crystal feeding the opa by a photon with @xmath415 oriented at 45@xmath24(t.h . ) , the other photon exciting @xmath416 _ without _ @xmath26 . the work by x. zou , l.wang and l. mandel in phys . 67 , 318 , ( 1991 ) is different from the present one since there no entanglement nor amplification effects were considered . c. su and k. wodkiewicz , phys . rev . a , 44 , 6097 ( 1991 ) . munro , m.d . reid , quantum opt.6 , 1 ( 1994 ) . owing to space limitations we postpone to another paper the analysis of a conceivable bell - type _ multiparticle _ nonlocality test by our system . f. de martini and y. r. shen , phys . rev . lett . 36 , 216 ( 1976 ) ; f.de martini , p. mataloni , e.palange,-y.r.shen , phys.rev.lett . 37 , 440 ( 1976 ) ; f.de martini , m.colocci , s.kohn and y.r.shen , phys . 38 , 1223(1977 ) . greenberger , m.a . horne and a. zeilinger , am . j. phys . 58 , 1131 ( 1990 ) ; n.d . mermin , phys.rev.lett . 65,1838 ( 1990 ) and : revs . physics , 65 , 803 ( 1993 ) ; a. cabello and g. alcaine , phys.rev.lett.80,1797 ( 1998 )
the new process of _ quantum - injection _ into an optical parametric amplifier operating in _ entangled _ configuration is adopted to _ amplify _ into a large dimensionality spin-@xmath0{0xbd}}% % beginexpansion { \frac12}% % endexpansion ~$]hilbert space the quantum entanglement and superposition properties of the photon - couples generated by parametric down - conversion . the structure of the wigner function and of the field s correlation functions shows a _ decoherence - free _ , multiphoton _ schroedinger - cat _ behaviour of the emitted field which is largely detectable against the squeezed - vacuum noise . furthermore , owing to its entanglement character , the system is found to exhibit multi - particle quantum nonseparability and bell - type nonlocality properties . these relevant quantum features are analyzed for several _ travelling - wave _ optical configurations implying different input quantum - injection schemes . ( pacs numbers : 03.65.bz , 03.67.-a , 42.50.-p , 89.70.+c ) .
at this conference on the metal - rich universe , a talk on damped lyman alpha ( dla ) systems may seem misplaced , since there is now a considerable literature on these absorption selected galaxies that describes them as metal - poor ( e.g. pettini 2004 and references therein ) . faced with the almost universally sub - solar abundances of dlas over all redshifts ( the average metallicity of dlas is @xmath1 at @xmath2 and @xmath3 and @xmath4 ) , questions have been raised concerning dla selection techniques . for example , are we missing a large fraction of metal - rich dlas due to dust obscuration bias ( ostriker & heisler 1984 ) ? or are the bulk of the metals in absorbers below the canonical dla column density threshold ( proux et al . 2006 ) ? in this contribution , i review the latest observations on the possibility of selection bias and discuss whether dlas can ever be considered as metal - rich . dla identification relies on a bright background qso on whose continuum the strong lyman @xmath5 and metal line species can imprint their absorption signature . if an intervening galaxy is rich in dust and metals , the very background qso on which we rely will appear fainter and redder and may ` drop out ' of traditional quasar surveys . this idea has been around for more than 20 years ( ostriker & heisler 1984 ; fall & pei 1993 ) and it has been suggested that extinction bias may ` hide ' more than 70% of gas at high redshift . four pieces of observational support are often cited for the dust bias scenario 1 ) only mild redshift evolution in dla metallicities , 2 ) low metallicities compared with emission line galaxies at similar redshifts , 3 ) steeper continuum slopes in qsos with dlas compared to those without ( pei , fall & bechtold 1991 ) and 4 ) an anti - correlation between n(hi ) and metallicity ( prantzos & boissier 2000 , see figure 1 ) . in order to quantify the impact of dust bias , two surveys based on radio - selected qsos have now been completed . together , the complete optical and radio absorption line system ( corals ; ellison et al . 2001 ) and the ucsd survey ( jorgenson et al . 2006 ) cover a redshift path @xmath6 for 119 radio - selected qsos with very deep ( and , in the case of corals , complete ) optical follow - up . neither survey finds an excess of absorbers either at high or low ( ellison et al . 2004 ) redshift and the neutral gas content is in good agreement ( within a factor of 2 ) with optical samples , see figure 2 . more importantly for the present discussion , akerman et al ( 2005 ) have shown that the metallicities of dlas in the corals sample are not significantly higher than optically selected samples and do not populate the parameter space at high n(hi ) and high metallicity ( figure 1 ) . however , since metallicities are usually weighted by the rare high n(hi ) absorbers , a larger sample is required in order to make this result robust . so , how do we explain the the observational ` evidence ' in support of dust bias ? a number of recent papers point to the idea that although dlas maybe reservoirs rich in atomic gas , they do not generally flag the location of the bulk of star - formation and , therefore , metals ( e.g. wolfe & chen 2006 ) . indeed , emission line spectroscopy of @xmath0 galaxies causing absorption line systems typically have solar abundances ( ellison , kewley & mallen - ornelas 2005 ) . such observations indicate that high abundances _ can _ be found in absorption selected galaxies , although they maybe confined to smaller regions than the cross - section of dla - producing gas . the cause of the anti - correlation between n(hi ) and metallicity is still under debate . however , ellison , hall & lira ( 2005 ) have argued that the very low values of reddening that are now being determined for dlas may soon make dust obscuration an unviable explanation . an alternative explanation may be that sightlines that pass through high column density , high metallicity gas , are simply rare . simulations support this idea ( e.g. johansson & efstathiou 2006 ) , showing that the cross - section for such gas at @xmath2 is small . finally , concerning the reddening of qso continua , this may simply have been a case of a difficult measurement combined with small number statistics . fitting continua to @xmath7 sdss spectra , murphy & liske ( 2004 ) have found reddening to be very low : e(b@xmath8v ) @xmath9 0.02 and a similarly low value has been found by ellison , hall & lira ( 2005 ) in corals qsos , e(b@xmath8v ) @xmath9 0.04 based on optical - to - ir colours . therefore , regardless of concerns about the modest sample size of radio - selected qso surveys such as ucsd and corals , there is no compelling observational evidence to invoke a dust bias . combined with this revised observational view of dust obscuration , it is interesting to note that theory is also re - assessing the effect of extinction . for example , trenti & stiavelli ( 2006 ) estimate the total gas density in dlas to be underestimated by only @xmath10% in optical surveys . although the majority of dlas are metal and dust poor , this is not to say that a small fraction of absorbers do not exhibit more extreme properties . these extreme cases open the door to some innovative analyses , permitting detection of unusual species and even shedding light on the galaxy s physical properties . diffuse interstellar bands ( dib ) are common in the spectra of reddened stars in the milky way . although numerous ( there are now over 100 different known bands ) , the identification of the dib carriers is one of the oldest outstanding mysteries in astronomical spectroscopy . amongst the potential candidates are polycyclic aromatic hydrocarbons and long carbon chains . dibs have been detected in only a handful of extra - galactic sightlines , including the lmc and smc ( see snow 2001 for a review ) and in one case , the broad 4428 feature in a @xmath0 dla ( junkkarinen et al 2004 ) . since the strength of some dib lines , such as the 5780 feature , correlate with n(hi ) in the galaxy ( herbig 1993 ) selecting high column density dlas may allow us to detect dibs in high redshift galaxies . we have been undertaking such a search ( lawton et al . in preparation ) and summarise our findings in figure 3 . in only one case do we detect dib absorption : in the @xmath0 dla towards ao 0235 + 164 ( york et al . 2006 ) where we detect both the 5780 and the 5705 dib . from this detection and upper limits from 4 other dlas , we find that the 5780 line strengths in dlas are weaker , by up to a factor of 10 , for their n(hi ) than galactic sightlines . similar deficiencies are seen in magellanic sightlines , suggesting that dib strength likely depends on metallicity and local physical properties as well as n(hi ) . on the other hand , the 5780 and 5705 dibs have similar equivalent width ratios in the @xmath0 dla and the galaxy , possible evidence that these two bands originate from a similar character . surprisingly , we do not detect the 6284 dib which is usually much stronger than the 5780 line which we do detect . this indicates that the physical conditions in the dla are different to the bulk of galactic and magellanic sightlines that have been studied . the only known sightline with similarly weak 6284 absorption is towards the sightline sk 143 , located in the smc wing . finally , in contrast to the trends with n(hi ) , galactic , magellanic and dla sightlines show a similar trend of e(b@xmath8v ) with dib strength . since dlas generally have low e(b@xmath8v ) ( ellison , hall & lira 2005 ) this implies that dibs are unlikely to be commonly detected in these absorbers . herbert - fort et al . ( 2006 ) have recently identified a sample of ` metal - strong ' absorbers from the sdss , characterised by strong heavy element absorption that is clearly detected even in low resolution sdss spectra . these metal - strong dlas comprise @xmath11 5% of the total population . in some cases , the very strong metal lines are a symptom of high n(hi ) . in other cases , the metallicity is truly high compared with known dlas . although relatively rare , the statistical power of sloan is expected to reveal several hundred metal - strong dlas . follow - up observations with high resolution spectrographs will yield data for a myriad of applications , including the search for rarely detected atomic transitions , molecular species and the study of isotopic ratios . in closing , it is interesting to note that although the definition of ` metal - rich ' at this conference has been somewhat subjective , with the exception of agn ( see e.g. max pettini s contribution to these proceedings ) we rarely find metallicities above @xmath12 in either individual stars , hii regions or in galaxies . however , a handful of very super - solar abundances ( up to @xmath13 ) have been recently reported for qso absorbers just below the traditional dla column density criterion of n(hi ) @xmath14 @xmath15 ( proux et al 2006 ; prochaska et al . the surprisingly high metallicities in these absorbers leads to a volley of new questions : how have these absorbers become so metal - rich at early times ? what are their low redshift analogues ? and what is the implication for the cosmic metals budget ? i am fortunate to enjoy stimulating collaborations with a large number of of colleagues . in particular for the work described here , chris akerman , chris churchill , pat hall , stephane herbert - fort , lisa kewley , brandon lawton , paulina lira , gabriela mallen - ornelas , max pettini , jason x. prochaska , ted snow and brian york . akerman , c. j. , ellison , s. l. , pettini , m. , steidel , c. c. , 2005 , _ a&a _ , * 440 * , 499 ellison , s. l. , churchill , c. w. , rix , s. a. , pettini , m. , 2004 , _ apj _ , * 615 * , 118 ellison , s. l. , hall , p. b. , lira , p. , 2005 , _ aj _ , * 130 * , 1345 ellison , s. l. , kewley , l. j. , mallen - ornelas , g. , 2005 , _ mnras _ , * 357 * , 354 ellison , s. l. , et al . , 2001 , _ a&a _ , * 379 * , 393 fall , s. m. , pei , y. c. , 1993 , _ apj _ , * 402 * , 479 herbert - fort , s. , et al . , 2006 , _ pasp _ , * 118 * , 1077 herbig , g. , 1993 , _ apj _ , * 407 * , 142 johansson , p. h. , efstathiou , g. , 2006 , _ mnras _ , * 371 * , 1519 jorgenson , r. a. , 2006 , _ apj _ , * 646 * , 730 junkkarinen , v. t. , 2004 , _ apj _ , * 614 * , 658 murphy , m. t. , liske , j. , 2004 , _ mnras _ , * 354 * , l31 ostriker , j. p. , heisler , j. , 1984 , _ apj _ , * 278 * , 10 pei , y. c. , fall , s. m. , bechtold , j. , 1991 , _ apj _ , * 378 * , 6 proux , c. , et al . , 2006 _ a&a _ , * 450 * , 53 pettini , m. , 2004 , proceedings of ` cosmochemistry . the melting pot of the elements ' , edited by c. esteban , r. j. garc lpez , a. herrero , f. snchez , 257 . prantzos , n. , boissier , s. , 2000 , _ mnras _ , * 315 * , 82 prochaska , j. x. , et al . , 2006 , _ apj _ , * 648 * , l97 snow , t. p. , 2001 , proceedings of gaseous matter in galaxies and intergalactic space , editted by r. ferlet , m. lemoine , j .- m . desert , b. raban , 63 trenti , m. , stiavelli , m. , 2006 , _ apj _ , * in press * , astro - ph/0606693 wolfe , a. m. , chen , h .- w . , 2006 , _ apj _ , * in press * , astro - ph/0608040 york , b. a. , et al . , 2006 , _ apj _ , * 647 * , l29
although damped lyman alpha ( dla ) systems are usually considered metal - poor , it has been suggested that this could be due to observational bias against metal - enriched absorbers . i review recent surveys to quantify the particular issue of dust obscuration bias and demonstrate that there is currently no compelling observational evidence to support a widespread effect due to extinction . on the other hand , a small sub - set of dlas may be metal - rich and i review some recent observations of these metal - rich absorbers and the detection of diffuse interstellar bands in one dla at @xmath0 .
let @xmath2 be a reductive lie algebra defined over an algebraically closed field of characteristic zero . assume that @xmath2 is @xmath3-graded @xmath4 with even part @xmath5 and odd part @xmath6 . we may refer to such a symmetric lie algebra as the _ symmetric pair _ @xmath7 . the algebraic adjoint group @xmath8 of @xmath2 acts on @xmath2 and the closed connected subgroup @xmath9 with lie algebra @xmath5 acts on @xmath6 . a sheet of @xmath6 ( resp . @xmath2 ) is an irreducible component of a locally closed set of the form @xmath10 sheets of @xmath2 have been extensively studied in several papers in the past decades . on one hand , the key papers of borho and kraft @xcite describe sheets as disjoint union of so - called _ decomposition classes _ ( also known as _ jordan classes _ ) . the decomposition classes form a finite partition of @xmath2 , each class being irreducible , locally closed and of constant orbit dimension . in particular , there is a dense class in each given sheet . an important notion used in @xcite is the parabolic induction of orbits introduced in @xcite which gives rise to a notion of induction of decomposition classes . it is shown in @xcite that a sheet @xmath11 with dense decomposition class @xmath12 is precisely the union of the decomposition classes induced from @xmath12 . two consequences of this are the following . firstly , each sheet contains a unique nilpotent orbit . secondly , there is a parameterization of sheets coming with induction @xcite . later , broer developed methods for a deeper study of decomposition classes , their closures , regular closures , quotients and normalizations in @xcite . on the other hand , making use of the previous parameterization , it is shown in @xcite that sheets are parameterized by their slodowy slices . more explicitly , let @xmath13 be a representative of the nilpotent orbit of a sheet @xmath11 of @xmath2 and embed @xmath13 in an @xmath14-triple @xmath15 . the slodowy slice of @xmath11 ( with respect to @xmath13 ) is @xmath16 where @xmath17 stands for the centraliser of @xmath18 in @xmath2 . katsylo proves that @xmath19 and that a geometric quotient of @xmath11 can be expressed as a finite quotient of @xmath20 . in @xcite , i m hof shows that the morphism @xmath21 is smooth . this relates smoothness of @xmath11 to smoothness of @xmath20 and eventually leads to a proof of smoothness of sheets in classical lie algebras . to our knowledge , the only known case of a singular sheet lies in a simple lie algebra of type g@xmath0 @xcite . in this case , the two non - trivial sheets ( _ i.e. _ non - regular and with more than one orbit ) are the two irreducible components of the set of subregular elements . one of these subregular sheets @xmath22 is smooth while the other @xmath23 is singular . more precisely , we can see that three analytical germs of @xmath23 intersect in the neighborhood of elements of the subregular nilpotent orbit . in @xcite , an explicit desingularization of @xmath23 is constructed in terms of the well - known projection @xmath24 . we now look at the symmetric case and sheets of @xmath6 . most of the ground results the authors are aware of in this setting are gathered in @xcite . an important feature is that there exists a notion of _ @xmath25-decomposition class _ in @xmath6 and these classes share several good properties with the usual decomposition classes of @xmath2 . in particular , there is still a unique dense @xmath25-decomposition class in each sheet of @xmath6 . in addition , each sheet of @xmath6 contains at least one nilpotent orbit . however the uniqueness statement no longer holds in general . one of the obstacles rising in the study of the sheets of @xmath6 is the lack of a well behaved notion of parabolic induction as one can check in the @xmath26-case . for instance the induction theory developed in @xcite does not preserve orbit codimension , hence is of little help for our purpose . the main philosophy of @xcite consists in noting that , at least in the case @xmath27 , the slodowy slices of a sheet @xmath11 of @xmath6 still seem to encode significant geometric information of @xmath11 . one of the aim of the present work is to justify this assertion in a more systematic manner . for instance , we show in section [ slosli ] that several properties of sheets , such as dimension , smoothness and orbits involved , are fully reflected in the corresponding slodowy slices . we state these results for wider classes of subvarieties of @xmath6 in proposition [ propsmooth ] , theorem [ genpart ] and proposition [ deformation1 ] . then we introduce in section [ sliceind ] the notion of slice induction . it turns out to be precise enough to rebuilt important parts of the theory resulting from parabolic induction in the lie algebra case . this includes ( see theorem [ thmind ] and corollary [ corind ] ) : construction of one parameter deformations of orbits . stratification properties for decomposition classes . characterization of a sheet with dense @xmath25-decomposition class @xmath12 as the union of induced classes from @xmath12 . in the lie algebra case , this also provides a new insight on these results . in @xcite , the connection between sheets of @xmath2 and sheets of @xmath6 was studied . in particular it was shown that , whenever sheets of @xmath2 are smooth , sheets of @xmath6 are also smooth . the last goal of this paper is the description of sheets of @xmath6 in the non - trivial symmetric lie algebra of type g@xmath0 . we make use of two approaches for this study . in section [ g2slsl ] , we study the subregular sheets through their slodowy slices . this provides the set - theoretical description of the sheets and describes the behavior of the singularities of @xmath23 through the intersection with @xmath6 . in section [ g24ality ] , we exploit the symmetry of @xmath2 making use of @xmath28-ality as described in @xcite . this allows us to construct a nice desingularization in @xmath29 of the singular sheet ( proposition [ desing ] ) , following the guidelines of @xcite . it is plausible that most of what is stated in section [ slosli ] and [ sliceind ] remains true in the more general setting of @xmath30-representation . however , the authors are unaware of references for a general theory of decomposition classes in this setting . we are grateful to the referees for the fast and careful reading of the paper . in particular , section [ sliceind ] has benefited from their remarks . we start with some notation . in the whole paper , @xmath31 is an algebraically closed field of characteristic @xmath32 , @xmath33 is a reductive symmetric lie algebra over @xmath31 , that is a @xmath3-graded reductive lie algebra @xmath2 with even part @xmath5 and odd part @xmath6 . in particular , @xmath5 is a lie algebra and @xmath6 is a @xmath5-module . we denote by @xmath34 the semisimple part of @xmath2 . lie algebras can be seen as particular cases of symmetric lie algebras in the following sense : given a lie algebra @xmath2 , there exists a symmetric lie algebra @xmath35 such that the @xmath36-module @xmath37 is isomorphic to the @xmath2-module @xmath2 , @xmath38 and @xmath39 . ] . as a consequence , all the statements enounced below in the symmetric setting hold for lie algebras replacing both @xmath5 and @xmath6 by @xmath2 , and @xmath25 by @xmath8 . a large part of the lie theory have a symmetric counterpart . we refer to @xcite for the ground results on symmetric lie algebra . let @xmath8 be the adjoint group of @xmath2 and @xmath25 be the closed connected subgroup of @xmath8 with lie algebra @xmath40 . the group @xmath25 acts on @xmath6 . for @xmath41 , it follows from ( * ? ? ? * proposition 5 ) that @xmath42 for @xmath43 , we set @xmath44 . note that we can replace @xmath8 by @xmath25 in the previous definition when @xmath45 thanks to . for @xmath46 , the centralizer in @xmath47 of @xmath48 is denoted by @xmath49=0\}$ ] . if @xmath50 is a @xmath3-subspace of @xmath2 we write @xmath51 , @xmath52 and we have @xmath53 we say that a levi subalgebra @xmath54 _ arises from @xmath6 _ if there exists a semisimple element @xmath55 such that @xmath56 ( it corresponds to the notion of subsymmetric pair in @xcite ) . in this case , @xmath57 and @xmath58 are reductive and semisimple @xmath3-graded lie subalgebras of @xmath2 . in addition , we can decompose @xmath2 in @xmath57-modules in the following way @xmath59 where @xmath60 is the orthogonal complement of @xmath57 in @xmath34 with respect to the killing form . more concretely , if @xmath56 we can write @xmath61 $ ] . it is well - known that @xmath62 is the center of @xmath57 . in particular , . moreover , we have @xcite @xmath64 we denote by @xmath65 ( resp . @xmath66 ) the closed connected subgroup of @xmath25 with lie algebra @xmath67 ( resp . @xmath68 ) . then , @xmath69 and the @xmath65-orbits of @xmath70 are precisely the orbits associated to the reductive symmetric lie algebra @xmath71 . the same holds for @xmath66-orbits ( = @xmath65-orbits ) of @xmath72 . define @xmath73 the next lemma shows that @xmath25-orbits and @xmath65-orbits of elements of @xmath74 are closely related . [ ts ] let @xmath57 be a levi subalgebra arising from @xmath6 . then , the following conditions are equivalent for any @xmath75 . a. @xmath76 , b. @xmath77 , where @xmath78 is the semisimple component of @xmath79 , c. @xmath80 , d. @xmath81={\mathfrak{p}}_{{\mathfrak{l}}^{\perp}}\oplus[{\mathfrak{k}}_{{\mathfrak{l}}},y]$ ] . moreover , @xmath74 is an open subset of @xmath70 . we have @xmath82 . hence @xmath83=[{\mathfrak{l}}^{\perp},y]\oplus[{\mathfrak{l}},y]\subset { \mathfrak{l}}^{\perp}\oplus [ { \mathfrak{l}},y]$ ] , with equality if and only if @xmath84 if and only if @xmath85 . as a consequence , ( i ) is equivalent to ( iv ) : @xmath83={\mathfrak{l}}^{\perp}\oplus [ { \mathfrak{l}},y]$ ] . by the way , we also note that @xmath84 is an open condition on @xmath79 . on the other hand , we see that ( iv ) is just ( iv ) intersected with @xmath6 . let @xmath55 such that @xmath56 . with the help of , we have @xmath86)-\dim [ { \mathfrak{k}},y]&=&\dim[{\mathfrak{k}},v]+\dim [ { \mathfrak{k}}_{{\mathfrak{l}}},y]-\dim [ { \mathfrak{k}},y]\\ & = & \frac 12(\dim[{\mathfrak{g}},v]+\dim [ { \mathfrak{l}},y]-\dim [ { \mathfrak{g}},y])\\ & = & \frac12(\dim ( { \mathfrak{l}}^{\perp}\oplus[{\mathfrak{l}},y])-\dim [ { \mathfrak{g}},y])\end{aligned}\ ] ] since we have the inclusion @xmath81\subset{\mathfrak{p}}_{{\mathfrak{l}}^{\perp}}\oplus[{\mathfrak{k}}_{{\mathfrak{l}}},y]$ ] , we get the equivalence between ( iv ) and ( iv ) . through tangeant spaces , we also see that ( iii ) is equivalent to ( iv ) . there remains to show that ( i ) is equivalent to ( ii ) . denote the nilpotent part of @xmath79 by @xmath87 . we have @xmath88 so ( ii ) implies ( i ) . let us now assume that @xmath89 is not included in @xmath57 . since @xmath90 , the endomorphism @xmath91 stabilizes the non - trivial subspace @xmath92 . since @xmath91 is nilpotent , there exists a non - zero element in @xmath93 . hence , @xmath94 . by contraposition , ( i ) implies ( ii ) . [ defxl]given a levi subalgebra @xmath57 of @xmath2 arising from @xmath6 , a nilpotent element @xmath95 embedded in a normal @xmath14-triple @xmath96 ( here _ normal _ means @xmath97 , @xmath98 and we allow @xmath99 as @xmath14-triple ) and a subset @xmath100 , we define @xmath101 via @xmath102 where @xmath74 is as in . we say that @xmath103 is the _ slodowy slice _ of @xmath12 with respect to @xmath104 . this is a generalization of the notion appearing in @xcite . in fact , in the case @xmath105 , we have @xmath106 and @xmath107 is the natural analogue in the symmetric setting of the ordinary slodowy slice in lie algebras , see @xcite . in what follows , a cone of @xmath6 means a subset of @xmath6 stable under multiplication by @xmath108 . the next lemma is well - known ( see _ e.g. _ @xcite ) . [ lmelimx ] let @xmath12 be a @xmath25-stable cone of @xmath6 and @xmath15 a normal @xmath14-triple such that @xmath109 . + then @xmath13 belongs to each irreducible component of @xmath110 . in particular , @xmath111 . a standard proof of this lemma is based on the construction of a one - parameter subgroup of @xmath112 which contracts @xmath113 to @xmath13 . define the characteristic grading @xmath114 by @xmath115=ix\}$ ] . for @xmath116 , let @xmath117 be such that @xmath118 we have @xmath119 and @xmath120 since @xmath121 is compatible with the characteristic grading . on the other hand , it is easy to show @xcite that @xmath122 , hence @xmath123 . as a consequence , @xmath124 normalizes the cone @xmath12 and hence @xmath125 . since @xmath126 , we have @xmath127 for any @xmath128 . since @xmath129 is a one parameter subgroup of @xmath130 , each irreducible component of @xmath113 is stable under the @xmath124-action and the lemma follows . next , we wish to enlight the strong connection linking @xmath12 and @xmath131 . this was lacking in @xcite in the general case and the following proposition renders some definitions and techniques of this paper obsolete . for example , it can easily be used together with @xcite ) to show that condition ( @xmath132 ) of @xcite is automatically satisfied . [ propsmooth ] let @xmath57 be a levi subalgebra of @xmath2 arising from @xmath6 and @xmath15 a normal @xmath14-triple of @xmath58 . let @xmath12 be an irreducible locally closed @xmath25-stable subset of @xmath6 such that @xmath133 and @xmath134 be a locally closed subset in @xmath135 . set @xmath136 , @xmath137 and @xmath138 . 1 . the action of @xmath25 on @xmath6 yields a smooth morphism @xmath139 of relative dimension @xmath140 . 2 . @xmath141 . 3 . @xmath142 is a pure locally closed variety of codimension @xmath143 in @xmath135 . 4 . let @xmath144 be an irreducible component of @xmath142 . then @xmath145 is dense in @xmath12 . @xmath146 restricts to a smooth dominant morphism @xmath147 first of all , @xmath146 is @xmath25-equivariant with respect to the @xmath25-action on the domain of @xmath146 given by @xmath148 . hence it is sufficient to check that @xmath146 is smooth at points of the form @xmath149 . moreover , since the domain and codomain of @xmath146 are smooth ( recall from lemma [ ts ] that @xmath150 is open in the affine space @xmath151 ) , it is sufficient to check that the induced map on the tangeant spaces is surjective ( * ? ? ? * vii remark 1.2 ) . at the point @xmath149 , it is given by @xmath152+x}$ ] . hence , thanks to lemma [ ts ] , we have @xmath153\oplus { \mathfrak{p}}_{{\mathfrak{l}}}^f={\mathfrak{p}}_{{\mathfrak{l}}^{\perp}}\oplus [ { \mathfrak{k}}_{{\mathfrak{l}}},y]\oplus{\mathfrak{p}}_{{\mathfrak{l}}}^f$ ] . this implies that @xmath154 is smooth at @xmath155 if and only if @xmath156 is , where @xmath157 is the morphism yielded by the action of @xmath65 on @xmath70 . in other words , we can restrict ourselves to the case @xmath158 . in this case , we follow ( * ? ? ? * corollary 1 ) . we easily see from graded @xmath14-theory that @xmath159\oplus{\mathfrak{p}}^f$ ] so @xmath146 is smooth at @xmath160 . consider the @xmath161-action on @xmath162 , given by @xmath163 . then @xmath146 is equivariant with respect to this action . hence the open set of smooth points of @xmath146 is stable under @xmath124 . therefore , it follows from that @xmath146 is smooth on @xmath164 and hence on the whole domain @xmath165 . + ( ii ) we have @xmath166 . assertion ( i ) implies that @xmath146 has constant fiber dimension . hence , the dimension of any fiber of @xmath167 is less or equal to @xmath140 and we have @xmath168 . + ( iii - iv ) ( * ? ? * i proposition 7.1 ) ( which also holds for locally closed subsets of an affine space ) states that @xmath169 for any irreducible component @xmath144 of @xmath142 . on the other hand , since @xmath12 is @xmath25-stable , we have @xmath170 and we deduce @xmath171 from ( ii ) . since @xmath12 is irreducible , this proves ( iii ) and ( iv ) . + ( v ) it follows from ( iv ) that @xmath172 is dominant . in order to obtain smoothness , we apply the argument of base extension by @xmath173 as in ( * ? ? ? * and above ) . more details can also be found in ( * ? ? ? * proposition 3.9 ) . [ propopen ] let @xmath174 be as in proposition [ propsmooth ] , omitting the assumption @xmath133 . then 1 . @xmath142 is a dense open subset of @xmath175 , 2 . @xmath176 is an open subset of @xmath12 . \(i ) if @xmath177 , there is nothing to prove . from now on , we assume that it is non - empty . since @xmath12 is open in @xmath178 , the subset @xmath142 is open in @xmath175 . let @xmath144 be an irreducible component of @xmath175 . then @xmath145 is a dense constructible subset of @xmath178 by propostion [ propsmooth](iv ) . hence it meets @xmath12 and , since @xmath12 is @xmath25-stable , we have @xmath179 . in other words , the open subset @xmath180 meets each irreducible components of @xmath175 . + ( ii ) since smooth morphisms are open ( * ? ? ? * v theorem 5.1 and vii theorem 1.8 ) , the result is a consequence of ( i ) and proposition [ propsmooth](v ) . the equivalent statements ( i ) , ( ii ) and ( iii ) of the following theorem are inspired by similar properties in the lie algebra case , which can be deduced from parabolic induction theory when @xmath12 is a decomposition class ( see _ e.g. _ @xcite ) . [ genpart ] let @xmath181 be a nilpotent element embedded in a normal @xmath14-triple @xmath182 and let @xmath12 be a locally closed @xmath25-stable cone . the following conditions are equivalent : 1 . @xmath111 . 2 . there exists a non - empty open set @xmath183 such that @xmath184 for any @xmath185 . there exists @xmath186 such that @xmath184 . 5 . @xmath187 . the implications ( ii)@xmath188(iii)@xmath188(i)@xmath188(v ) are obvious and ( v)@xmath188(iv ) is a consequence of proposition [ propopen](i ) applied to an irreducible component of @xmath178 meeting @xmath189 . + let us now prove ( iv)@xmath188(ii ) . take @xmath190 . it is open in j by proposition [ propopen](ii ) and , for any @xmath185 , we have @xmath191 . then , our implication is a consequence of lemma [ lmelimx ] applied to the @xmath25-stable cone @xmath192 . from the computational point of view , the previous theorem may be of key importance . indeed , the existence of a degeneration from @xmath12 to @xmath13 reduces to the existence of an element in the intersection of @xmath12 with the affine space @xmath189 , which might be much easier to check . on the other hand , we have the following proposition derived from ( * ? ? ? * theorem 38.6.9 ( i ) ) . [ deformation0 ] if @xmath12 is a @xmath25-stable cone of @xmath193 , then there exists a nilpotent element @xmath194 such that 1 . @xmath111 , 2 . @xmath195 . [ deformation1 ] let @xmath196 be a locally closed @xmath25-stable cone and @xmath197 a set of representatives of the nilpotent @xmath25-orbits satisfying ( i ) and ( ii ) of proposition [ deformation0 ] embedded in normal @xmath14-triples @xmath198 . let @xmath199 . + then , @xmath200 is an open cover of @xmath12 . the sets @xmath201 are open thanks to proposition [ propopen](ii ) . pick @xmath186 , then it follows from proposition [ deformation0 ] that there exists @xmath202 such that @xmath203 . applying theorem [ genpart ] to @xmath192 , we get @xmath204 , so @xmath205 . since @xmath201 is stable under the action of @xmath112 , we get @xmath206 . [ smoothequiv ] 1 . a major consequence of this proposition is that the whole geometry of @xmath12 is closely related to the geometry of its different slodowy slices . indeed , locally , we can assume that @xmath207 for some @xmath202 and it follows from proposition [ propsmooth](v ) that this variety is smoothly equivalent to @xmath208 . this also provides a more solid ground to the philosophy of @xcite , where it is proven in some particular cases that the slodowy slices contains enough information to describe the whole variety . the main drawback of this approach is that it does not yield @xmath209 when @xmath12 is irreducible in the lie algebra case , contrary to other parameterization such as @xcite . however , this drawback is somehow necessary since the property @xmath209 may fail in the symmetric case ( see _ e.g. _ @xcite for the description of the regular sheet when @xmath210 ) . we have several examples in mind of such locally closed @xmath25-stable cone included in some @xmath211 . we have already seen that @xmath192 plays a role in the previous proofs . we can also consider a @xmath25-decomposition class @xmath212 , or its regular closure @xmath213 . sheets are particular examples of the last type . recall from the beginning of section [ slosli ] that a levi subalgebra @xmath57 is said to _ arise from @xmath6 _ if it is the centralizer of a semisimple element @xmath214 . recall also that , in this case , the orbits in @xmath70 associated to the symmetric lie algebra structure @xmath71 coincide with the @xmath215-orbits . a _ datum _ of @xmath7 is a pair @xmath216 where @xmath57 is a levi subalgebra of @xmath2 arising from @xmath6 and @xmath217 is a nilpotent @xmath65-orbit in @xmath70 . the _ @xmath25-decomposition class _ defined by a datum @xmath216 is @xmath218 [ rqdec ] 1 . these definitions make sense with @xmath7 replaced by any other reductive other reductive symmetric pair . for instance , if @xmath219 is a levi subalgebra arising from @xmath6 and @xmath220 is a datum of @xmath221 , we have @xmath222 where @xmath223 is a regularity condition taken with respect to the action of @xmath224 . the classes we have defined are the symmetric analogue of decomposition classes studied in @xcite in the lie algebra case . we call these last classes _ @xmath8-decomposition classes _ and they are of the form @xmath225 with @xmath226 any levi subalgebra and @xmath217 a nilpotent @xmath227-orbit in @xmath57 . in fact , using notation of section [ slosli ] , the @xmath8-decomposition classes are exactly the images of the @xmath228-decomposition classes under the natural isomorphism @xmath229 . these are of the form @xmath230 with @xmath231 and @xmath232 . the connection between a @xmath25-decomposition class @xmath12 and the @xmath8-decomposition class containing @xmath12 is studied in @xcite . 3 . the @xmath25-orbits of levi - factors arising from @xmath6 are in one to one correspondance with their @xmath8-orbits which are , in turn , easily characterized through the satake diagram of @xmath7 , cf . @xcite . hence , in order to have a convenient classification of @xmath25-decomposition classes , it would be enough to understand in which cases @xmath233 is @xmath234-conjugate to @xmath235 . this plays an important role in the classification of sheets when @xmath236 , @xmath237 are rigid orbits as shown in @xcite . however , for our purposes , we will not need such a classification . the @xmath25-decomposition classes are equivalence classes of elements sharing a similar jordan decomposition in the following sense : @xmath238 and we refer to @xcite for most of the known properties of these classes ( called @xmath25-jordan classes in _ loc . _ ) in the symmetric lie algebra case . in particular , we note that @xmath25-decomposition classes form a finite partition of @xmath6 . these classes are also irreducible , locally closed and of constant orbit dimension . as a consequence , in each sheet @xmath11 there exists a dense open @xmath25-decomposition class @xmath239 and we have @xmath240 in the lie algebra case , such regular closures of @xmath8-decomposition classes have been studied in @xcite , where an important parameterization is given in theorem 5.4 ( see ( * ? ? ? * lemma 3.2 ) for a refinement ) . in @xcite , this understanding has been deepened thanks to the parabolic induction theory of nilpotent orbits . let us recall the definition . take two levi subalgebras @xmath241 , with corresponding subgroups @xmath242 , @xmath243 . one can choose a parabolic subalgebra @xmath244 of @xmath245 with levi - factor @xmath246 . denote by @xmath247 the nilpotent radical of @xmath244 . given a nilpotent @xmath248-orbit @xmath249 , _ the _ parabolically induced orbit from @xmath246 to @xmath245 is the dense nilpotent @xmath250-orbit @xmath251 of @xmath252 . for the purpose of our exposition , we will say in such case that the pair @xmath253 ( resp . the @xmath8-decomposition class @xmath254 ) _ parabolically induces _ @xmath255 ( resp . @xmath256 ) . the aim of this section is to extend a significant number of the known properties of @xmath8-decomposition classes of @xcite to @xmath25-decomposition classes . however , as mentioned in the introduction , the parabolic subalgebras behave badly in general for a symmetric lie algebra . this is why we introduce a new notion of induction , using the tools developed in section [ slosli ] . this notion turns out to coincide with the parabolic induction in the lie algebra case , see corollary [ corind ] . [ definduction ] 1 . [ weakind ] given two data @xmath257 ( @xmath237 ) of @xmath7 , we say that @xmath220 _ weakly slice induces _ @xmath255 if @xmath258 and @xmath259 where @xmath260 is an element of @xmath251 embedded in a normal @xmath14-triple @xmath261 of @xmath245 . [ mainind ] we say that @xmath220 _ slice induces _ @xmath255 if , moreover , @xmath262 for some @xmath263 . [ jordind ] given two @xmath25-decomposition classes @xmath264 ( @xmath237 ) , we say that @xmath212 ( weakly ) slice induces @xmath265 if there exist data @xmath257 with @xmath266 , ( @xmath237 ) such that @xmath220 ( weakly ) slice induces @xmath255 . when , there is no context of parabolic induction , the term _ induction _ will always refer to slice induction . when @xmath212 and @xmath265 are @xmath25-decomposition classes , we write @xmath267 [ rkdefind ] 1 . the definitions do not depend on the choices of @xmath260 or @xmath268 since all of these are @xmath224-conjugate . also , all the elements of @xmath269 share the same @xmath224-orbit dimension . finally , it is worth noting that the notions of induction and weak induction of data do only depend on the involved data and not on the ambient lie algebra @xmath2 . 2 . if @xmath270 , condition [ condind ] can be written in the nicer form : @xmath271 3 . assume that @xmath258 . the data @xmath220 ( weakly ) induces @xmath272 if and only if @xmath273 ( weakly ) induces @xmath274 . indeed , @xmath275 and @xmath276 . in particular , when studying induction , one can always assume that the ambient lie algebra @xmath2 is semisimple . assume that @xmath2 is semisimple . in view of theorem [ genpart ] and , a @xmath25-decomposition class @xmath212 weakly induces a nilpotent orbit @xmath217 ( _ i.e. _ the @xmath25-decomposition class @xmath277 ) if and only if @xmath278 . the purpose of the following theorem is to extend this property to induction of arbitrary @xmath25-decomposition classes . recall that @xmath142 is introduced in definition [ defxl ] . [ thmind ] let @xmath279 be @xmath25-decomposition classes . the following conditions are equivalent . a. [ wi ] @xmath280 , b. [ wii ] @xmath281 where @xmath272 is some ( any ) datum defining @xmath265 and @xmath268 is a normal @xmath14-triple in @xmath245 whose nilpositive element belongs to @xmath251 . c. [ wiii ] @xmath282 , d. [ wiv ] @xmath283 , @xmath188 is obvious . @xmath188 : choose @xmath284 and embed it in a normal @xmath14-triple @xmath268 of @xmath245 . since @xmath285 is @xmath25-stable we can choose an element @xmath286 in this intersection such that the jordan decomposition of @xmath286 is @xmath287 with @xmath288 and @xmath284 . therefore @xmath289 so @xmath290 and @xmath291 . then proposition [ propopen](i ) yields @xmath281 . @xmath188 : choose an element @xmath292 and consider its jordan decomposition @xmath293 . since @xmath294 , the centralizer @xmath295 is a levi subalgebra of @xmath245 ( lemma [ ts ] ) . hence it makes sense to speak of the @xmath224-decomposition class @xmath296 with @xmath297 . since @xmath298 , we derive from that @xmath299 belongs to this class . since @xmath300 , the datum @xmath220 weakly induces @xmath255 . @xmath188 : choose a datum @xmath301 defining @xmath212 , which weakly induces a datum @xmath302 defining @xmath265 . since @xmath303 , we have @xmath304 and @xmath305 where @xmath306 is the @xmath307-decomposition class @xmath308 . on the other hand , we have seen in remark [ rkdefind ] that our hypothesis implies that @xmath306 weakly induces @xmath309 in @xmath310 . then it follows from theorem [ genpart ] , applied to the symmetric pair @xmath311 , that @xmath312 . hence @xmath313 and we get @xmath314 by @xmath25-stability of @xmath315 . a key point in the proof of @xmath188 is that whenever we have @xmath316 with jordan decomposition @xmath317 , we can manage to realize the degeneration through a one parameter family @xmath318 , @xmath319 with the @xmath320 lying in a slodowy slice centered on @xmath321 and included in @xmath89 . hence the whole degeneration takes place in the levi @xmath89 and thus yields a degeneration toward any element with jordan decomposition similar to @xmath321 . [ dimind ] let @xmath212 , @xmath265 be @xmath25-decomposition classes such that @xmath280 . then @xmath322 if and only if the dimension of @xmath25-orbits in @xmath212 and @xmath265 coincide . choose a datum @xmath255 defining @xmath265 . we can argue as in theorem [ thmind ] ( ii)@xmath188(i ) to find a data @xmath220 defining @xmath212 and an element @xmath323 . on the other hand , any element @xmath324 , @xmath284 satisfies @xmath325 . then , it follows from lemma [ ts](iii ) that @xmath299 and @xmath286 share the same @xmath25-orbit codimension in @xmath6 if and only if they share the same @xmath224-orbit codimension in @xmath326 . since @xmath327 , the result follows . as a consequence of theorem [ thmind ] and lemma [ dimind ] , we are now able to prove the following . [ corind ] a. if @xmath239 is a @xmath25-decomposition class , then @xmath328 b. if @xmath212 and @xmath265 are @xmath25-decomposition classes , @xmath329 is a union of @xmath25-decomposition classes . c. sheets are union of @xmath25-decomposition classes . d. induction ( resp . weak induction ) of data and of @xmath25-decomposition classes are transitive . that is , for classes , @xmath330 @xmath331 e. [ equivind ] in the lie algebra case , the parabolic induction coincides with the slice induction . since @xmath25-decomposition classes form a partition of @xmath6 , the first part of ( i ) is an immediate consequence of the equivalence @xmath332 @xmath332 in theorem [ thmind ] . the second part of ( i ) follows from the first one and lemma [ dimind ] . then we deduce from ( i ) the statement ( iv ) concerning @xmath25-decomposition classes . let us now show ( iv ) on data . choose three data @xmath257 , @xmath333 such that @xmath257 ( weakly ) induces @xmath334 , @xmath237 . without loss of generality , we can assume that @xmath335 ( see remark [ rkdefind ] ) . then @xmath336 ( weakly ) induces @xmath337 for @xmath237 , so @xmath338 ( weakly ) induces @xmath339 . hence there exist data @xmath340 , @xmath25-conjugate to @xmath341 ( @xmath342 , and such that @xmath343 ( weakly ) induces @xmath344 . since @xmath345 , we have @xmath346 so , up to @xmath25-conjugacy , we can assume that @xmath347 . in order to show ( ii ) , we only need to note that @xmath329 is the union of @xmath25-decomposition classes induced both by @xmath212 and @xmath265 . statement ( iii ) follows from the fact that any sheet is the regular closure of a @xmath25-decomposition class . in @xcite it is shown that , whenever @xmath239 is a @xmath8-decomposition class , @xmath348 is the set of @xmath8-decomposition classes parabolically induced by @xmath239 . then for @xmath8-decomposition classes follows from ( i ) and the discussion in remark [ rqdec ] . we now prove for data , using arguments similar to the proof of ( iv ) for data . without loss of generality , we can restrict to the cases where a datum @xmath220 slice ( resp . parabolically ) induces a datum of the form @xmath349 . since @xmath349 is the only datum defining @xmath350 , such a slice ( resp . parabolic ) induction is equivalent to the corresponding induction from @xmath351 to @xmath350 . the result on data then follows from the result on decomposition classes . an important consequence of corollary [ corind ] is that the slice induction can be seen as a generalization of the parabolic induction , fitting to the symmetric lie algebra setting needs . note in particular that ( i ) , ( ii ) , ( iv ) and ( iii ) are respective analogues of @xcite and @xcite . the fact that a closure of a decomposition class is a union of @xmath25-decomposition classes has also been shown in @xcite when the ground field is precisely @xmath352 . using the results of section [ slosli ] and of this section , we can reconsider the example given in @xcite ) ( where there is a mistake in the definition of @xmath57 ) . here @xmath353 , the sheet @xmath354 of @xmath2 is defined by @xmath355 with @xmath57 arising from @xmath6 of type @xmath356 ( long root ) . we know that @xmath357 is the single @xmath25-decomposition class @xmath358 ( * ? ? * theorem 7.8 ) . in particular , @xmath359 has only one irreducible component of maximal dimension @xmath360 , namely @xmath348 ( * ? ? ? * lemma 8.2 ) . any nilpotent element @xmath361 embedded in a normal @xmath14-triple @xmath362 satisfies @xmath363 if and only if @xmath364 ( see theorem [ genpart ] and proposition [ propsmooth ] ) . we choose the two nilpotent elements @xmath365 as in @xcite ) and embed them in @xmath14-triples @xmath366 . it turns out that the slodowy slices @xmath367 and @xmath368 are irreducible varieties of respective dimension @xmath369 and @xmath370 . in particular @xmath371 while @xmath372 . an other consequence is that @xmath359 has at least two irreducible components , @xmath348 of dimension @xmath373 and @xmath374 of dimension @xmath375 , see proposition [ deformation1 ] . this is a counterexample to the equidimensionality of @xmath376 for @xmath11 a sheet of @xmath2 , this property being observed when @xmath27 ( * ? ? ? * theorem 13.2 ) in this section , we assume that @xmath2 is a simple lie algebra of type g@xmath0 . we are interested in studying non - trivial ( i.e. non - regular and not restricted to one nilpotent orbit ) sheets in this case . we adopt conventions and notations of @xcite . in particular , we fix a cartan subalgebra @xmath377 of @xmath2 . in the corresponding root system , we fix a basis @xmath378 , with @xmath379 a short root , and label the associated positive roots as pictured we choose , as a chevalley basis , the one given in @xcite and denote it by @xmath380 , @xmath381 , @xmath382\!])$ ] , @xmath383\!])$ ] , with @xmath384 , @xmath385 and @xmath386 an @xmath387-triplefor @xmath237 . the nilpotent cone of @xmath2 consists of five nilpotent conjugacy classes . we denote the orbit of dimension @xmath388 ( @xmath389 ) by @xmath390 . moreover , we choose some particular representatives @xmath391 ( two of them when @xmath392 ) of these orbits as given in column 2 of table [ representants ] . since @xmath2 is of rank two , the only non - regular non - nilpotent elements of @xmath2 are semisimple subregular elements . this gives rise to two subregular sheets of @xmath2 which can be described as union of @xmath8-decomposition classes as follows . @xmath394 this follows from the fact that @xmath395 and @xmath396 are of the same dimension ( here , @xmath397 ) and that both must contain a nilpotent orbit in their regular closure ( proposition [ deformation0 ] ) . a practical criterion to distinguish generic elements of @xmath398 from those of @xmath399 is that elements of the former lie in a levi of type @xmath400 ( short root ) while their centraliser is a levi of type @xmath401 ( long root ) . it is known @xcite that @xmath398 is smooth at points of @xmath395 and is singular at points of @xmath402 , undergoing a threefold covering in a resolution of singularities , see _ e.g. _ proposition [ detg ] . on the other hand , @xmath399 is a smooth variety . let us describe the @xmath25-orbits of interest in this setting . first , the symmetric lie algebra is of maximal rank , hence any @xmath8-orbit intersects @xmath6 @xcite . since semisimple orbits intersect @xmath6 into single orbits ( see _ e.g. _ ( * ? ? ? * proposition 6.6 ) ) , there are exactly two subregular semisimple @xmath25-decomposition classes : @xmath406 , with @xmath407 an element of @xmath408 , @xmath237 . it turns out @xcite that @xmath409 is a single @xmath25-orbit @xmath410 for @xmath411 and is the union of two @xmath25-orbits @xmath412 and @xmath413 in the subregular case @xmath392 ( respectively numbered by @xmath414 and @xmath28 in @xcite ) . we refer to table [ representants ] for representatives of these different @xmath25-orbits . note that , since sheets of @xmath6 are union of @xmath25-decomposition classes , there are two 6-dimensional sheets of subregular elements . these are @xmath415 and @xmath416 . even if each of these sheets must contain at least a nilpotent orbit among @xmath412 and @xmath413 ( proposition [ deformation0 ] ) , it is a non - trivial problem to decide which orbits belong to a given sheet . at this point , we may even not rule out the existence of a possible @xmath417-dimensional subregular sheet consisting of a single orbit @xmath412 or @xmath413 . a. the only sheets of subregular elements in @xmath6 are @xmath418 and @xmath419 . b. the decomposition of each sheet @xmath420 ( @xmath237 ) as union of @xmath25-decomposition classes is @xmath421 c. the sheet @xmath418 is smooth on @xmath212 and @xmath412 . at points of @xmath413 , the singularities of @xmath418 are smoothly equivalent to the intersection of three lines . d. the sheet @xmath419 is smooth . the statements ( ii ) , ( iii ) and ( iv ) rely on computations on the slodowy slices . in fact , embedding @xmath422 and @xmath423 into respective normal @xmath14-triple , @xmath424 and @xmath425 , we claim that in particular , @xmath435 so @xmath264 slice induces @xmath436 for @xmath437 and @xmath438 . hence proposition [ descsheetg2 ] ( ii ) follows from corollary [ corind ] ( or , in a more simple way , from lemma [ lmelimx ] ) . the other consequence of lemma [ exg2 ] is that @xmath439 ( @xmath237 ) ( resp . @xmath440 ) is smoothly equivalent to the ( smooth ) affine line @xmath441 ( resp . @xmath442 ) , see remark [ smoothequiv ] . on the other hand @xmath443 is smoothly equivalement to a union of three lines meeting at a point . this explains ( iii ) and ( iv ) of proposition [ descsheetg2 ] . to sum up , the picture is the following . only one branch in the neighborhood of the singularities of @xmath23 at points of @xmath412 is preserved under the intersection with @xmath6 . on the contrary , the singularities at points of @xmath413 are `` intact '' when intersected with @xmath6 . the idea of the proof of lemma [ exg2 ] is the following . due to lemma [ ft ] , an element in the slodowy slice is subregular if and only if it is not regular . in particular , we get explicit equations of the subregular locus in the slodowy slice via minors of some matrices . then , these equations are simple enough to solve . most of the computations have been made by hand and then checked by @xcite using w. de graaf s package @xcite . we give the explicit outcome of the computation in terms of our chevalley basis . set @xmath444 and @xmath445 . then @xmath446 and @xmath447 are normal @xmath14-triples and the elements @xmath448 of the lemma can be chosen as follows . 1 . note that , thanks to , our method also provides explicit one - parameter degenerations @xmath450 ( @xmath438 ) 2 . in the real case , we have no reason to think that decomposition classes closure inclusions always leaves footprints on the slodowy slice level ( see proof of the crucial statement proposition [ propsmooth](iii ) ) . however , it is an interesting feature to note that our method may still provide real degeneration in some cases . indeed , here , all the @xmath448 belong to the obvious real form of @xmath2 . 3 . the computations presented in this section can be improved in order to get similar results in higher rank . for example , in a work in progress , the first author has found two singular sheets in the ( ordinary ) lie algebra of type f@xmath451 , the others being smooth . the singular sheets are those appearing in line 9 and 10 in the last table of @xcite . we now provide an explicit desingularization of @xmath418 through the well - known projection @xmath24 . we first note that it would be hard to follow broer s approach @xcite in the lie algebra case for normalization and desingularization since it uses fiber bundles of the form @xmath452 with @xmath453 a parabolic subgroup of @xmath8 . in order to achieve our goal , we use an other description of our algebra @xmath2 of type g@xmath0 using @xmath28-ality as defined in @xcite . namely , we choose four copies @xmath454 of a @xmath370-dimensional space equipped with a non - degenerate bilinear skew - symmetric form @xmath455 . then @xmath456 can be equipped with a lie bracket in such a way that @xmath457 , see for an identification . in this model , @xmath458 is a lie subalgebra of @xmath459 and acts on @xmath460 in the usual way . we refer to @xcite for the definition of the bracket of two elements of @xmath461 using @xmath455 which , for instance , identifies @xmath462 with its dual and hence @xmath463 with @xmath464 . this presentation of @xmath465 relies on the @xmath466-symmetry of @xmath465 as can be seen on the extended dynkin diagram of type @xmath467 . we are interested in the @xmath468 ( resp . @xmath469)-action on @xmath459 induced by permutations on the @xmath370 spaces @xmath462 and @xmath470 ( resp . on the @xmath414 spaces @xmath471 ) . its fixed point space @xmath472 ( resp . @xmath473 ) is isomorphic to @xmath29 ( resp . @xmath2 ) and can be described as @xmath474 where @xmath475 and @xmath476 are again copies of our two - dimensional space . choosing a cartan subalgebra of @xmath477 in @xmath478 and weight vectors @xmath479 ( resp . @xmath480 ) in @xmath475 ( resp . @xmath481 ) , we rediscover the combinatorics of the root spaces in @xmath477 as pictured below . = [ scale=0.7 ] ( 0,0 ) ( 1,0 ) node[right]@xmath482 ; ( 0,0 ) ( -3/2,0.866 ) node[left]@xmath483 ; ( 0,0 ) ( -1/2,0.866 ) node[above]@xmath484 ; ( 0,0 ) ( 1/2,0.866 ) node[above ] ; ( 0,0 ) ( 3/2,0.866 ) node[right]@xmath485 ; ( 0,0 ) ( 0,1.732 ) node[above]@xmath486 ; ( 0,0 ) ( -1,0 ) ; ( 0,0 ) ( 3/2,-0.866 ) node[right]@xmath487 ; ( 0,0 ) ( 1/2,-0.866 ) node[below ] ; ( 0,0 ) ( -1/2,-0.866 ) node[below]@xmath488 ; ( 0,0 ) ( -3/2,-0.866 ) node[left]@xmath489 ; ( 0,0 ) ( 0,-1.732 ) ; ( 0,0 ) ellipse ( 8ex and 1ex ) ; ( 0,0 ) ellipse ( 1ex and 12ex ) ; moreover , we have a natural projection @xmath490 , given for instance on @xmath491 via @xmath492 . its kernel is a @xmath493-module so @xmath494 is @xmath495-equivariant . we also introduce faithful representations that allow us to manipulate more easily these algebras . let @xmath496 with action of @xmath461 given by @xmath497 and standard action of @xmath498 . it is one of the three inequivalent fundamental representations of @xmath457 of dimension @xmath499 and @xmath500 is the subalgebra preserving the symmetric form @xmath501 as an @xmath502-module , @xmath503 decomposes as @xmath504 where @xmath505 , @xmath506 and @xmath507 . setting @xmath508 , one gets that @xmath509 is the subalgebra of @xmath465 stabilizing both subspaces @xmath510 and @xmath511 while @xmath512 is the subalgebra of @xmath509 stabilizing both subspaces @xmath513 and @xmath514 . in particular , we recover that @xmath515 and that @xmath516 is a symmetric pair isomorphic to @xmath517 @xcite . we denote by @xmath518 the @xmath519-orbit of dimension @xmath373 in @xmath509 . its young diagram is ( 3,1,1,1,1 ) . moreover , we let @xmath520 be the set elements of @xmath509 of rank at most @xmath370 with respect to the representation on @xmath511 . the closed set @xmath520 is nothing but the union @xmath521 where @xmath522 is any semisimple element whose centraliser is a levi of type b@xmath0 . in particular @xmath523 is a sheet in @xmath509 . we refer to @xcite for the classification of nilpotent orbits in @xmath491 . in @xmath509 , the @xmath519-orbit @xmath528 splits into two @xmath529-orbits @xmath530 and @xmath531 whose respective @xmath532-diagrams are ( aba , a , b , b , b ) and ( bab , a , a , b , b ) . we also define @xmath533 . we can then state the following result \(ii ) an element of @xmath491 acts on @xmath503 in the following way : @xmath538 \hline \scriptstyle{0 } & 0 & p\\ \hline \scriptstyle{0 } & p^{\vee } & \phantom{a } 0 \rule{0pt}{15pt}\phantom{a}\end{array}\right ) \begin{array}{l}\scriptscriptstyle{v_1}\\v_a\\v_b\rule{0pt}{15pt } \end{array}\ ] ] where @xmath539 and @xmath540 is the dual of @xmath541 with respect to @xmath542 . we are interested into rank @xmath370 elements of this form and these are exactly elements for which @xmath541 is of rank @xmath369 . we make use of the @xmath28-ality setting to describe such elements . an element of @xmath543 induces a rank one element of @xmath544 ( and , equivalently , of @xmath545 ) if and only if it is a pure tensor of the form @xmath546 with @xmath547 and @xmath548 . after computation , one gets table [ fibers ] where the column @xmath549 presents the @xmath25-decomposition class of @xmath550 under consideration . up to easy @xmath495-conjugation , one can turn our representatives of table [ representants ] to the element @xmath551 in column @xmath552 . finally , column @xmath553 present all elements in the fiber @xmath554 . @xmath555{*}{${\mathcal{o}}_4 $ } & \multirow{3}[-3]{*}{$c_+^2c_-\otimes d_+$ } & c_{+}^2\otimes c_-\otimes d_+\\ \cline{4 - 4 } & & c_+c_-\otimes c_+\otimes d_+\\ \hline { \mathcal{o}}_3 & c_+^3\otimes d_+&c_+^2\otimes c_+\otimes d_+\\ \hline 0 & 0&0\\ \(i ) the desingularization property follows from @xmath556 for @xmath557 and the fact that @xmath558 is a smooth variety . let us check this last point . under the natural isomorphism @xmath559 , we have @xmath560 and @xmath561 . the smoothness is then a consequence of the transitivity of the action of @xmath562 on @xmath558 . in view of proposition [ detg ] , there only remains to see that @xmath563 for @xmath438 . for instance , since @xmath564 are isotropic vectors with respect to @xmath455 , we can check that the pre - image @xmath565 of @xmath422 sends @xmath566 on a non - zero multiple of @xmath567 and this last one is sent on a non - zero multiple of @xmath568 . hence the @xmath532-diagram of @xmath569 is ( aba , a , b , b , b ) and @xmath570 . 999 a. altman and s kleiman , _ introduction to grothendieck duality theory _ , lecture notes in math . , * 146 * , springer - verlag , 1970 . l. v. antonyan , on classification of homogeneous elements of @xmath571-graded semisimple lie algebras , _ vestnik moskov . _ , * 37 * ( 1982 ) , 29 - 34 . w. borho , ber schichten halbeinfacher lie - algebren , _ invent . math . _ , * 65 * ( 1981 ) , 283 - 317 . w. borho and h. kraft , ber bahnen und deren deformationen bei linear aktionen reduktiver gruppen , _ comment . math _ , * 54 * ( 1979 ) , 61 - 104 . a. broer , decomposition varieties in semisimple lie algebras , _ canad . j. math . _ , * 50 * ( 1998 ) , 929 - 971 . a. broer , lectures on decomposition classes . _ in : _ representation theories and algebraic geometry ( montreal , pq , 1997 ) , _ nato adv . _ , * 514 * , kluwer acad . , 1998 , 39 - 83 . m. bulois , sheets of symmetric lie algebras and slodowy slices , _ j. lie theory _ , * 21 * ( 2011 ) , 1 - 54 . d. z. djokovic , classification of nilpotent elements in simple exceptional real lie algebras of inner type and description of their centralizers , _ j. algebra _ , * 112 * ( 1988 ) , 503 - 524 . w. fulton and j. harris , _ representation theory _ , graduate texts in mathematics , * 129 * , springer - verlag , 1991 . the gap group , _ gap groups , algorithms , and programming , version 4.6.4 _ ; 2013 , ( http://www.gap-system.org ) r. hartshorne , _ algebraic geometry _ , graduate texts in mathematics , * 52 * , springer - verlag , 1977 . p. hivert , nappes sous - rgulires et quations de certaines compactifications magnifiques , http://tel.archives-ouvertes.fr/docs/00/56/45/94/pdf/these.pdf , ph . d thesis , universit de versailles ( 2010 ) . [ in french ] a. e. i m hof , the sheets of classical lie algebra , http://pages.unibas.ch/diss/2005/dissb_7184.pdf , ph . d. thesis , university of basel ( 2005 ) . i. katsylo , sections of sheets in a reductive algebraic lie algebra , _ math . ussr - izv . _ , * 20 * ( 1983 ) , 449 - 458 . b. kostant and s. rallis , orbits and representations associated with symmetric spaces , _ amer . _ , * 93 * ( 1971 ) , 753 - 809 . h. kraft , closures of conjugacy classes in @xmath572 , _ j. algebra _ , * 126 * ( 1989 ) , 454 - 465 . j. landsberg and l. manivel , representation theory and projective geometry . _ in _ : algebraic transformation groups and algebraic varieties , _ enc . math . _ , * 132 * , springer - verlag , 2004 , 131 - 167 . m. le barbier grnewald , the variety of reductions for a reductive symmetric pair , _ transform . groups _ , * 16 * ( 2011 ) , 1 - 26 . t. levasseur and s. p. smith , primitive ideals and nilpotent orbits in type @xmath572 , _ j. algebra _ , * 114 * ( 1988 ) , 81 - 105 . g. lusztig and n. spaltenstein , induced unipotent classes , _ j. london math . ( 2 ) _ , * 19 * ( 1979 ) , 41 - 52 . t. ohta , the singularities of the closure of nilpotent orbits in certain symmetric pairs , _ tohoku math . j. _ , * 38 * ( 1986 ) , 441 - 468 . t. ohta , induction of nilpotent orbits for real reductive groups and associated varieties of standard representations , _ hiroshima math . * 29 * ( 1999 ) , 347 - 360 . d. i. panyushev and o. yakimova , symmetric pairs and associated commuting varieties , _ math . _ , * 143 * ( 2007 ) , 307 - 321 . d. peterson , geometry of the adjoint representation of a complex semisimple lie algebra , ph . d. thesis , harvard university ( 1978 ) . v. l. popov and e. a. tevelev , self - dual projective algebraic varieties associated with symmetric spaces . _ in _ : algebraic transformation groups and algebraic varieties , _ enc . math . _ , * 132 * , springer - verlag , 2004 , 131 - 167 . w. de graaf , sla - a gap package , 0.13 , 2013 , http://www.science.unitn.it/~degraaf/sla.html p. slodowy , simple singularities and simple algebraic groups , _ lecture notes in math . _ , * 815 * , springer - verlag , 1980 . p. tauvel and r. w. t. yu , _ lie algebras and algebraic groups _ , springer monographs in mathematics , springer - verlag , 2005 .
in this paper , we study sheets of symmetric lie algebras through their slodowy slices . in particular , we introduce a notion of _ slice induction _ of nilpotent orbits which coincides with the parabolic induction in the lie algebra case . we also study in more details the sheets of the non - trivial symmetric lie algebra of type g@xmath0 . we characterize their singular loci and provide a nice desingularization lying in @xmath1 .
the phase diagram of the quark matter has been actively investigated for decades @xcite . quarks are confined inside hadrons and can not be observed as free particles at low energy . on the other hand at high energy , quarks become free particles due to the asymptotic freedom of the strong interaction . therefore , it is expected that quarks undergo the phase transition between confined and deconfined states which is one of the most important issues in the theoretical and experimental particle physics . the fundamental theory to describe quark matter is quantum chromodynamics ( qcd ) , the theory of strong interaction . it is , however , not practical to extract reliable predictions at low energy due to the necessity of complicated nonperturbative calculations in this area . for this reason some effective approaches are used such as the nambu jona - lasinio ( njl ) model @xcite and its polyakov - loop incorporated version , the pnjl model @xcite , the linear sigma model @xcite , the chiral perturbation theory @xcite , the lattice qcd simulations @xcite . in this letter , we will consider the njl model known as a low - energy effective theory of qcd ( for reviews , see , @xcite ) . at low temperature , @xmath0 , and chemical potential , @xmath1 , constituent quarks are heavy due to the chiral symmetry spontaneous breaking while they are expected to be light at high @xmath0 and/or @xmath1 where the chiral symmetry is getting restored . thus the quark system is closely related to the phenomenon of the chiral phase transition . the njl model actually predicts the chiral symmetry breaking at low energy and its restoration at high energy . many investigations of the phase diagram are based on the njl and pnjl models ( see , e.g. , @xcite and ) . since the njl model is not renormalizable , the model predictions inevitably depend on a regularization procedure applied . the most frequently used method is probably the three - momentum cutoff regularization which introduces the cutoff scale @xmath2 . the model in the cutoff scheme may miss an important contribution when the quark density becomes comparable to the cutoff scale . there is an alternative method , the dimensional regularization ( dr ) , to avoid the issue @xcite . in the dr , divergences coming from fermion loop integrals are regularized by lowering the dimension of the integration through an analytic continuation in the dimension variable . using various regularization ways is interesting , because we believe that the regularization scheme is a dynamical part of the njl model , it is related to the size and shape of the effective quark interaction as discussed in @xcite . it was found that the model with the dr nicely describes quark systems at low energy , such characteristics as the phase structure and meson properties @xcite . we shall study in this article the phase diagram in the three flavor njl model with the dr . it is interesting because the recent work by the present authors @xcite indicates that the phase structure , especially the order of the transition , may differ drastically from the one in the cutoff regularization . the structure of this letter is following : in sec . [ njl_model ] , the three flavor njl model and its parameters are presented . [ sec_critical ] is devoted to the explanation on the procedure of drawing the phase diagram . we then display the resulting phase diagram of the model in sec . [ sec_pd ] . the concluding remarks are given in sec . [ conclusion ] . the lagrangian of the three flavor model is @xmath3 , \label{l_4 } \\ \mathcal{l}_6 \!\!\!\ ! & = & \!\!\ ! -k \left [ \det\overline{q}_i ( 1-\gamma_5 ) q_j + { \rm h.c.\ } \right ] . \label{l_6}\end{aligned}\ ] ] where @xmath4 represents the diagonal mass matrix @xmath5 with flavor indices @xmath6 . @xmath7 and @xmath8 are the four- and six - fermion couplings , @xmath9 are the gell - mann matrices in flavor space with @xmath10 . the determinant in @xmath11 runs over flavor space , so this leads to the six - point interaction known as kobayashi - maskawa t hooft ( kmt ) term @xcite . the vacuum of the model is determined by the minimum of the thermodynamic potential @xmath12 with the partition function @xmath13 , the inverse temperature @xmath14 , and the volume of the system @xmath15 . applying the mean - field approximation , we can calculate the potential @xmath16 in the imaginary time formalism , @xmath17 , \label{omega0}\\ \omega_t \!\!\!\ ! & = & \!\!\ ! -\frac{2^{d/2}n_c } { 2 } t \int \!\ ! \frac{d^{d-1}p}{(2\pi)^{d-1 } } \sum_{i,\,\pm } \ln \bigl[1 + e^{-\beta e_i^{\pm } } \bigr ] . \label{omegat}\end{aligned}\ ] ] here @xmath18 corresponds to the vacuum contribution by the chiral condensates , @xmath19 and @xmath20 denote the temperature independent and dependent contributions , @xmath21 is the chiral condensate for each quark which is the order parameter of the model , @xmath22 is the number of colors . @xmath23 denotes dimensions in the fermion loop integral , @xmath24 is the energy of the quasi - particle with the constituent quark mass @xmath25 , @xmath26 with a quark chemical potential @xmath27 ) . the fermion loop integral in eq.([omega0 ] ) diverges , therefore we will perform the analytic continuation in @xmath23 to regularize it by decreasing the dimension @xmath23 as discussed in @xcite . in the cutoff scheme , the divergent contribution is dropped by introducing the momentum cutoff @xmath2 . to be more precise , the regularization in the dr and cutoff schemes are performed by the following replacements @xmath28 } m_0^{4-d } \!\!\ ! \int_0^{\infty } \!\!dp \,p^{d-2 } , \\ \int \!\!\frac{d^{d-1}p}{(2\pi)^{d-1 } } \!\!\!\ ! & \rightarrow & \!\!\!\ ! \frac{1}{2\pi^2 } \int_0^{\lambda } \!\!dp \,p^{2},\end{aligned}\ ] ] where @xmath29 is the renormalization scale which is needed to render physical quantities correct mass dimensions . as mentioned in the introduction , the constituent quark mass @xmath30 is closely related to the chiral symmetry breaking , namely to the value of @xmath31 . the self - consistent gap equations ( [ gap_eq ] ) are obtained as the condition for the thermodynamic potential to be at the extremum , @xmath32 . eqs ( [ gap_eq ] ) explicitly show that the difference between constituent and current quark masses is due to the underlying chiral symmetry breaking . the njl model with the dimensional regularization has 7 free parameters : current quark mass @xmath33 , @xmath34 , @xmath35 , the four- and six - point couplings @xmath7 , @xmath8 , the dimension @xmath23 , and the renormalization scale @xmath29 . we consider , for simplicity , the isospin symmetric case , @xmath36 , and set several values for @xmath37 . we then fix the remaining parameters by choosing 5 physical quantities among listed below : @xmath38 following @xcite , we name the parameter sets as case @xmath39 and @xmath40 depending on which quantities are selected . the case @xmath39 ( @xmath40 ) is fitted by @xmath41 . the parameter setting was performed in @xcite , and we shall employ three parameter sets , case @xmath42 , @xmath40 and @xmath39 , which are shown in tabs . [ para_etald ] , [ para_eta ] and [ para_chi170 ] . .case @xmath39 . [ cols="<,^,^,^,^,^",options="header " , ] in this section we explain how to draw the phase diagram of the model through the analysis of the thermodynamical potential and the gap equations . a critical temperature @xmath43 or chemical potential @xmath44 are given by the maxima of @xmath45 in fact we apply @xmath46 for low @xmath1 ( @xmath0 ) in crossover region . the above quantity becomes infinite at @xmath47 ) when the transition is of the first order . in this case we determine the transition boundary by the point where the discontinuous change of the chiral condensate @xmath48 occurs by directly searching the minimum of the thermodynamic potential . it is obvious that this procedure is consistent with the criterion of eq . ( [ dpdt ] ) , because a divergent point coincides the maximum point . to see the tendency of the phase transition , we show the behavior of @xmath49 for the case @xmath40 and cutoff with @xmath50mev near the transition boundary in fig . [ thermo ] . along the line @xmath51 in the case @xmath40 with @xmath50mev for @xmath52 , @xmath53 , @xmath54mev and @xmath55mev . right panel : @xmath56 along the line @xmath57 in the case cutoff with @xmath50mev for @xmath52 , @xmath58 , @xmath59mev and @xmath60mev . the circles indicate the global minima.,title="fig:",width=288 ] along the line @xmath51 in the case @xmath40 with @xmath50mev for @xmath52 , @xmath53 , @xmath54mev and @xmath55mev . right panel : @xmath56 along the line @xmath57 in the case cutoff with @xmath50mev for @xmath52 , @xmath58 , @xmath59mev and @xmath60mev . the circles indicate the global minima.,title="fig:",width=288 ] the curves are plotted along the line @xmath51 for @xmath52 , @xmath53 and @xmath54mev with @xmath55mev in the left panel , and along the line @xmath57 for @xmath52 , @xmath58 , @xmath59mev with @xmath60mev in the right panel . these lines are chosen so as to show the global minima for lower @xmath61mev and higher @xmath62mev , which are indicated by the circles , near the transition temperature @xmath63mev . @xmath64 denotes the chiral condensate @xmath48 at @xmath65 for each case . the striking difference is observed between these figures . there exists a bump between two stable minima in the dr case , which means that the transition is of the first order between @xmath66 and @xmath54mev . on the other hand the cutoff case ( right panel ) produces rather monotonous curves with no bump , which leads to a smooth crossover . the difference stems from the fact that the ratio of the thermal contribution ( @xmath1 dependence ) @xmath67 in the dr case is larger than that in the cutoff case at low @xmath0 . thus we confirm the stronger tendency of the first order phase transition in the dr scheme . in the crossover region , it is technically easier to analyze eq . ( [ dpdt ] ) through solving the gap equations because @xmath48 changes continuously with respect to @xmath69 . we show the numerical results in fig . [ fig_dpdt ] . in the case @xmath40 with @xmath50mev.,width=288 ] one sees that the maximum point moves toward lower @xmath0 with increasing @xmath1 , and the peak becomes large at high @xmath1 . the peak actually diverges when @xmath0 and @xmath1 coincide with the critical point @xmath70 . below @xmath71 , the transition becomes of the first order , and the analysis by eq . ( [ dpdt ] ) is no longer practically useful for the determination of the transition boundary as mentioned above . we are now ready to discuss the phase structure of the njl model with the dr . [ pd_etald ] displays the typical structure of the phase diagram in the model with the dr in the case @xmath42 . with @xmath72mev . the solid ( dashed ) line represents the first order ( crossover ) transition . the circle indicates the critical point.,width=288 ] this is a reasonable picture of a system in the chiral symmetry broken phase at low @xmath0 and @xmath1 , and in the chiral symmetry restored phase at high @xmath0 and/or @xmath1 . the solid ( dashed ) line represents the first order ( crossover ) transition , and the circle indicates the critical point . note that the transition temperature , @xmath73mev for @xmath74 , is comparable with the lattice qcd prediction , @xmath75mev . the critical point is located at @xmath76 , and it is interesting to see that @xmath71 is close to one obtained in the pnjl model with the cutoff regularization , @xmath77mev , for frequently used parameter set of @xcite , whereas @xmath78mev in the njl model fukushima:2008wg . note that the obtained critical point is close to one obtained in a njl type model with the smooth form factor @xcite , @xmath79 , and in the linear sigma model @xcite , @xmath80 . below we make more detailed comparison between the dr and the cutoff schemes . [ pd ] shows the phase diagrams in the cases @xmath40 and @xmath39 for various @xmath33 . and @xmath39 are shown in the left and right panels . the solid ( dashed ) lines represent the first order ( crossover ) transition . the circles indicate the critical points.,title="fig:",width=288 ] and @xmath39 are shown in the left and right panels . the solid ( dashed ) lines represent the first order ( crossover ) transition . the circles indicate the critical points.,title="fig:",width=288 ] we note that in the case @xmath40 , the region of chiral symmetry broken phase becomes smaller with choosing the smaller value of @xmath33 . on the other hand the case @xmath39 produces similar curves for different @xmath33 . the different behavior can be explained by the fact that the constituent quark mass @xmath81 gets smaller with decreasing @xmath33 in the case @xmath40 , while it almost does not change in the case @xmath39 as discussed in @xcite . in the cutoff case ( fig . [ pd_cut ] ) the region of the chiral symmetry broken state shrinks when @xmath33 is lowered as observed in the case @xmath40 . it is very interesting to note that the critical point disappears below @xmath82mev , where the transition is crossover for all @xmath0 and @xmath1 . a striking difference between the two regularizations is in that the critical point moves towards higher temperature with decreasing @xmath33 in the dr , while it moves to the opposite direction in the cutoff case . the difference may be understood by observing the value of the six - point coupling @xmath8 which becomes larger ( smaller ) with decreasing @xmath33 in the dr ( cutoff ) procedure , since the kmt term shown in eq.([l_6 ] ) tends to drive the first order phase transition @xcite . as discussed in @xcite , the constituent quark masses undergo two discontinuous changes at low @xmath0 in the dr scheme . and @xmath84 for @xmath52mev in the case @xmath42 with @xmath72mev.,width=288 ] fig . [ am_etald ] displays the typical behavior of @xmath85 and @xmath84 as functions of @xmath1 at low @xmath0 ( @xmath86mev ) , plotted in the case @xmath42 with @xmath72mev . one clearly observes two gaps : one is located around @xmath87mev and the other is around @xmath88mev . here we call these discontinuities as first and second gaps for lower and higher chemical potential , respectively . the first gap comes from the effect of the approximate @xmath89 restoration and the second one comes from that of the partial @xmath90 restoration . thus it may be interesting to study the phase structure concerning the second transition as well . to draw the phase diagram on the second transition , we set the criterion of the transition by using the following quantity @xmath91 then below @xmath92 , namely in the crossover region , the above quantity has only one maximum , which determines the crossover transition on @xmath83 . while above @xmath92 the quantity @xmath93 shows non - trivial behavior ; it becomes infinite at @xmath94 , and has second maximum at @xmath95 . so @xmath93 has typical two maxima at @xmath94 and @xmath95 below @xmath71 as seen in fig . [ am_etald ] . here we call the transition point corresponding to the second maximum , @xmath95 , the second phase boundary " . to distinguish between the two phase transitions , we call the transition line on @xmath48 discussed in the previous subsection the first phase boundary " . in the phase diagram on the 1st and 2nd phase boundaries ( fig . [ pd_etald_2nd ] ) with @xmath72mev . the red dashed and blue dotted lines represent the crossover transition on @xmath48 and @xmath83 . the red and blue solid lines indicate the transition on the first and second gap , respectively . the circles exhibit the critical points.,width=288 ] the dashed and dotted lines represent the crossover transition on @xmath48 and @xmath83 , respectively . the solid line for lower ( higher ) chemical potential indicates the discontinuous change on the first ( second ) gap . we see that the crossover line on @xmath83 is observed at a bit lower temperature than that on @xmath48 for @xmath96 . it should be noticed that the critical curves on @xmath48 and @xmath83 intersect at the critical end point ( @xmath97 ) on @xmath48 . because the value of @xmath83 is affected by @xmath48 , as is clearly seen from fig . [ am_etald ] , @xmath83 shows discontinuous change at the point where @xmath48 has a gap . then @xmath98 blows up and approaches to infinity near the critical point where @xmath99 is divergent . below @xmath71 , @xmath100 has two maxima appearing at the first gap and higher chemical potential . the first maximum coincides with the red solid line and the second one is plotted by the blue line in fig . [ pd_etald_2nd ] . the transition on the second gap also has the critical point whose location is exhibited by the blue circle at higher chemical potential . we also studied the other cases , @xmath40 and @xmath39 , with various @xmath33 , and found that the qualitative behavior does not show remarkable difference ; the critical point on @xmath83 moves toward higher temperature with decreasing @xmath33 as seen in the @xmath48 case . therefore , we only displayed the case @xmath42 here . we studied the phase diagram of the njl model with the dr and cutoff regularization . we found that the phase diagram on the @xmath101 plane in the njl model with the dr shows a stronger tendency of the first order phase transition . the tendency is confirmed by the shapes of the thermodynamic potential shown in fig . [ thermo ] where we find a bump in the dr , and rather monotonous behaviour in the cutoff case . we have also studied the phase structure on the change of @xmath83 in sec . [ subsec_phi_s ] . , where we found that the approximate @xmath89 symmetry and the partial @xmath90 symmetry restore at a similar temperature for low chemical potential , @xmath102 . it may be difficult to distinguish between the two lines experimentally , because the transitions are smooth crossover at low chemical potential . from the obtained phase diagrams , we conclude that the first order phase transition persists for low @xmath33 in the model with the dr method . the finding is consistent with the current symmetry analysis based consensus @xcite stating that the chiral phase transition is of the first order in the chiral limit , @xmath103 . this tendency may be understood by the following reasoning . the loop contribution from the lower integration momenta is enhanced by lowering dimension . it introduces non - locality in the model with the dr . the infrared behavior of the loop integral is important for thermal corrections . it can rise the critical end point temperature , @xmath71 . finally , because the parameter difference crucially affects the location of the critical point as confirmed in this article , we think it is interesting to study the related issues , such as the case with the chiral limit , and the @xmath104 dependence on the order of the chiral transition in the context of the columbia plot @xcite . hk is supported by the grant nsc-99 - 2811-m-033 - 017 from national science council ( nsc ) of taiwan . for recent reviews , m. a. stephanov , prog . * 153 * ( 2004 ) 139 , int . j. mod . phys . a * 20 * ( 2005 ) 4387 , pos lat * 2006 * ( 2006 ) 024 , o. philipsen , prog . * 174 * ( 2008 ) 206 , arxiv:1111.5370 [ hep - ph ] , w. weise , prog . * 186 * ( 2010 ) 390 , arxiv:1201.0950 [ nucl - th ] , k. fukushima and t. hatsuda , rept . * 74 * ( 2011 ) 014001 , g. endrodi , z. fodor , s. d. katz and k. k. szabo , jhep * 1104 * ( 2011 ) 001 . k. fukushima , phys . d * 77 * ( 2008 ) 114028 , erratum - ibid . d * 78 * ( 2008 ) 039902 . t. hell , s. roessner , m. cristoforetti and w. weise , phys . d * 79 * ( 2009 ) 014022 , phys . d * 81 * ( 2010 ) 074034 . t. inagaki , d. kimura , h. kohyama , and a. kvinikhidze , phys . d * 83 * ( 2011 ) 034005 . m. kobayashi and t. maskawa , prog . theor . phys . * 44 * ( 1970 ) 1422 , m. kobayashi , h. kondo , and t. maskawa , prog . * 45 * ( 1971 ) 1955 , g. t hooft , phys . d * 14 * ( 1976 ) 3432 , * 18 * ( 1978 ) 2199(e ) , phys . rept . * 142 * ( 1986 ) 357 .
we investigate the phase diagram on temperature - chemical potential plane in the nambu jona - lasinio model with the dimensional regularization . while the structure of the resulting diagram shows resemblance to the one in the frequently used cutoff regularization , some results of our study indicate striking difference between these regularizations . the diagram in the dimensional regularization exhibits strong tendency of the first order phase transition . qcd phase diagram , chiral symmetry , effective field theory
the ultraviolet - optical spectral energy distribution ( sed ) of quasars is usually called the `` big blue bump '' and characterised by a smooth continuum , which is considered to be thermal emission from an accretion disk , emission lines attributed to this emission reprocessed in surrounding clumped gas in a form of broad and narrow line regions as well as in a dust obscured torus , broad absorption lines ( possessed by about 10% of quasars ) related to some streams outward quasar centre , and sets of narrow absorption lines caused by intergalactic medium . despite of general similarity in uv - optical sed of quasars , they differ in equivalent width ( ew ) of emission lines and spectral index @xmath0 . the proximity of the regions producing emission in continuum and absorption lines ( i.e. the torus and the clumped broad line region , blr ) is considered to be the most promising explanation of the @xcite effect : the inverse correlation of ew of some emission lines with the monochromatic luminosities at uv region . this effect was found , e.g. , for ly@xmath2 ( 1215 ) , siii ( 1260 ) , oi ( 1304 ) , siiv+oiv ] ( 1400 ) lines ( see e.g. @xcite ) . but the physical explanation of the difference in @xmath0 and relation of quasar luminosity and lines ew to @xmath0 are still not clear . previously , using composite spectra of quasars with different @xmath0 , we showed that there is no correlation between @xmath0 and monochromatic luminosity at 1450 , @xmath1 ( @xcite ) . we also have not found any dependence of ew of lines within 12101450 on spectral index @xmath0 ( calculated within the range 12701480 ) ( @xcite ) . but due to the fact , that the latter study was carried out without separation by luminosity , its result requires verification on a sample of spectra with different luminosity as well . such verification has been done in the present study . we also have checked the possible relation between the spectral index and the mass of the central black hole . we used the sample of 3535 individual quasar spectra from the sloan digital sky survey ( sdss ) data release 7 with medium resolution of @xmath3 , compiled by @xcite . figure [ fig : distr ] shows the redshift and the spectral index distribution of this sample . we selected @xmath4 subsamples with different spectral index @xmath0 ( 16 steps , from @xmath5 to @xmath6 ) and monochromatic luminosity at 1450 , @xmath1 ( 12 steps , from 42.4 to 43.4 ) , and then compiled 192 composite spectra from them . here @xmath7 is in ergs@xmath8@xmath8 . figure [ fig : af ] shows the mean redshift , average logarithm of monochromatic luminosities at 1450 of subsamples and spectral index of composite spectra diagrams . figure [ fig:3-lum ] shows examples of composite spectra bunches with the same spectral index and different luminosity and vice versa , correspondingly . c c c the wavelength ranges 12151285 , 12901320 , 13201350 and 13501430 were considered separately and fitted with the help of the ` idl lmfit ` subroutine with a superposition of constant ( @xmath9 ) or power - law ( @xmath10 ) continuum and the smallest possible number of gaussian - profile emission lines in the following form : @xmath11,\ ] ] where @xmath12 is the rest - frame wavelength , @xmath13 , @xmath14 , @xmath15 are the amplitude , the central wavelength and the full width at half maximum ( fwhm ) ( up to @xmath16-factor ) of the @xmath17-th line . then the equivalent widths of these four separate sets of lines were calculated using obtained model parameters . figure [ fig : comp ] shows the whole wavelength range considered in the present study along with the known lines . c the virial mass of central supermassive black hole for each of 3535 individual quasar spectra was calculated using c iv emission line and the following empirical formula from @xcite : @xmath18 where @xmath19 here @xmath20 and @xmath21 are flux and luminosity , @xmath22 is the fwhm of c iv ( 1549 ) , @xmath23 is the photometric distance . we used the following values of cosmological parameters : @xmath24 kms@xmath8mpc@xmath8 , @xmath25 , @xmath26 from planck collaboration results ( @xcite ) , and calibration parameters @xmath27 and @xmath28 for c iv from @xcite . the obtained @xmath0ew diagrams for different luminosities are shown in figure [ fig : res1 ] . one can clearly see dependence between these two characteristics in some cases . in all cases the significance of possible dependence was checked with the f - test using constant and linear models as restricted and unrestricted ones , respectively . the resulting coefficients of the fitted models ( the constant @xmath29 for the first case and @xmath30 , @xmath31 for the second one ) along with their errors , and the values of cumulative probability @xmath32 are presented in tables 1 , 2 . summarising the obtained results gives us the following : \(i ) for lines ly@xmath2+ov+nv+siii@xmath33+siii ( 12151285 ) the ew decreases with @xmath0 ; the significance of the linear coefficient of regression in most cases is @xmath34 , and the baldwin effect is clearly seen for this set of lines . the more detailed analysis with separation of this region into three parts showed the same trend for ly@xmath2+ov and siii , unlike nv+siii@xmath33 for which we did not found any dependence . at the same time , the baldwin effect is seen for ly@xmath2+ov and siii lines , while for nv+siii@xmath33 we do not see any correlation between ew and luminosity , which agree with previous studies . presence of the baldwin effect was found by other authors for our first two parts together ( ly@xmath2+ov+nv+siii@xmath33 , see e.g. @xcite ) . some authors , e.g. @xcite tried to separate this set into two parts . all of them report on the baldwin effect for ly@xmath2+ov , and only @xcite claims for baldwin effect for nv+siii@xmath33 unlike two others who did not found baldwin effect for this spectral feature . \(ii ) for lines siiii@xmath33+oi+siii ( 12901320 ) the significance of the linear coefficient of regression differs for subsamples with different luminosity , moreover the values of @xmath31 differ in sign and have errors @xmath35 , thus we can not speak about some general trend . this confirms the absence of any dependence of ew on @xmath0 seen by eye . we also did not see any baldwin effect for this set of lines , although it was found by @xcite for oi line . \(iii ) unlike the first case , for two other sets of lines , cii+ov+caii ( 12901350 ) and x1+siiv+oiv]+x2 ( 12901350 ) , we found the increase of ew with @xmath0 , but only for the second set we can clearly see the baldwin effect , found previously by other authors . on the other hand , note that the significance of the @xmath0ew dependence in most cases are @xmath34 , but in the second case for some subsamples we have negative @xmath31 values and also large errors . .coefficients of the fitted models : the constant @xmath29 for the constant fit and @xmath30 , @xmath31 for the linear ( @xmath36 ) , and the values of cumulative probability @xmath32 for f - test . [ cols="^,^,^,^,^,^,^,^,^ " , ] figure [ fig : mass ] shows @xmath0m@xmath37 diagrams for composite spectra ( with different luminosities ) on the left and for all 3553 individual quasars on the right . it is clearly seen that there is no dependence in both cases . unlike the case of relation between the monochromatic luminosity of the quasar and the ew of its emission lines ( the baldwin effect ) , interpretation of relation between the spectral index , @xmath0 , and the ew is not so simple . firstly , one have to remember that the term ` spectral index ' describes here only a slope of some limited part of the uv - optical quasar sed ( the big blue bump ) . it has nothing common with a real power - law spectra inherent to non - thermal radiation , and serves here only as tool for some classification of quasar seds . the big blue bump is considered to be a result of thermal emission of inhomogeneous gas - dust torus , i.e. a superposition of a number of planck curves with different temperature . it has been shown previously that there is no correlation between @xmath0 and @xmath1 ( @xcite ) and in the present work we also show that there is no dependence between @xmath0 and the bh mass . hence we can assume that the shape of uv - optical sed of quasars does not depend , at least directly , on the central bh . it has to be influenced by the physical properties of the torus and can also be affected by the host galaxy . on the other hand , the obtained dependence between @xmath0 and ew of some emission lines , mostly for those lines for which the baldwin effect is observed , claims for some relation between the physical properties of the torus and the blr , e.g. the chemical abundance . dietrich , m. , hamann , f. , shields , j. c. , constantin , a. , vestergaard , m. , chaffee , f. , foltz , c. b. & junkkarinen , v. t. continuum and emission - line strength relations for a large active galactic nuclei sample , _ _ , 581 , 912 - 924 , 2002 . shang , z. , wills , b. j. , robinson , e. l. , wills , d. , laor , a. , xie , b. & yuan , j. the baldwin effect and black hole accretion : a spectral principal component analysis of a complete quasar sample , _ _ , 586 , 52 - 71 , 2003 . tang b. , shang z. , gu q. , brotherton m. s. , runnoe j. c. the optical and ultraviolet emission - line properties of bright quasars with detailed spectral energy distributions _ astr . j. s. s. _ , 201 , 38 , 2012 . torbaniuk , o. , ivashchenko , g. & sergijenko , o. in : j. safrankova & j. pavlu ( eds . ) , some spectral properties of the quasar ultraviolet bump _ wds12 proc . of contributed pap . : part iii physics _ ( prague : matfyzpress ) , 123 - 128 , 2012 .
using 192 composite spectra stacked from subsamples of individual sdss dr7 quasar spectra binned in spectral index , @xmath0 , and logarithm of monochromatic luminosity at 1450 , @xmath1 , we found that there is a dependence of the emission line equivalent width ( ew ) on spectral index ( correlation or anti - correlation ) for some lines , mostly for those lines for which the baldwin effect is detected . we also show that there is no dependence between the virial mass of the central supermassive black hole of a quasar and its spectral index @xmath0 .
we have undertaken an ir imaging campaign for the low - redshift 3crr ( * ? ? ? * ( laing , riley & longair 1983 ) ) radio sources using the irac ( @xmath1 , @xmath2 , @xmath3 , and @xmath4 band ) and mips ( @xmath5 , @xmath6 , and @xmath7 band ) cameras on the _ spitzer _ satellite ( * ? ? ? * ; * ? ? ? * ( fazio 2004 ; rieke 2004 ) ) . our main aims were ( 1 ) to quantify the extent to which signatures of the radio activity of the cores appear in the core ir emission ; ( 2 ) to investigate the heating of dust and gas in the core caused by the active nucleus ; ( 3 ) to investigate whether the host galaxies display any extended dust features , such as those that might arise from recent mergers or from uplifts from the core ; and ( 4 ) to image ir emission from radio structures such as jets and hot spots , and so to study the relativistic electron populations that they contain . the sensitivity and angular resolution that the spitzer cameras provide is a strong function of wavelength , and the images of the 3crr galaxies usually show structure in the host galaxy only in the irac bands , where the fwhm of the point spread function ( psf ) varies from @xmath8 to @xmath9 . the @xmath10 to @xmath11-arcsec psfs of the mips images provide structural information only for the closest objects , such as m87 ( @xmath12 3c274 ) , but since the ir emission is usually concentrated in the cores in the mips bands , the uncertainty in the core flux densities introduced by our inability to make precise fits to the structures of the galaxies is not usually a limiting factor in measuring useful core spectra . the basic reduction of the _ spitzer _ data used ` iracproc ` ( * ? ? ? * ( schuster , marengo & patten 2006 ) ) and mopex ( * ? ? ? * ( makovoz & marleau 2005 ) ) , and detailed image analysis and photometry were then performed using codes developed in bristol . photometry for point - like components is accurate to 5% to 20% , depending on band , and photometric errors represented in figures [ fig : irac_colours][fig : ir_spectra ] below include both systematic and statistical components . these errors become large in the irac bands at the high - redshift limit of the study , @xmath13 , because of poor structural discrimination between the cores and the host galaxies . the errors may also become large for faint emission in crowded fields , such as in fitting for the ir emission from the hot spots in 3c33 ( * ? ? ? * ( kraft 2007 ) ) . the irac structure of each host galaxy was fitted by an elliptical srsic model ( * ? ? ? * ( srsic 1963 ) ) with a point - like component at its centre , convolved with a model psf , and with foreground stars , jets , and overlapping galaxy images masked . this fitting procedure could also be used for the 24-@xmath14 m mips band in some cases . for most of the mips images , however , a sufficiently good fit was obtained with a simple point source model because of the relatively large psf . the signal / noise of the mips flux densities in the 70- and 160-@xmath14 m bands is sometimes quite poor because of the lumpy background structure in the images . some residual images from the fitting showed clear signs of jets , for example in 3c31 ( * ? ? ? * ( lanz 2011 ) ) , 3c66b ( * ? ? ? * ( earlier reported by tansley 2000 from iso data ) ) , and ngc6251 . in the case of ngc6251 the jet is seen to extend well beyond the optical image of the galaxy , and even beyond the irac frame . when the residual images were compared with the radio structures other features sometimes became apparent , such as ir - bright hotspots in cases such as 3c33 ( * ? ? ? * ( kraft 2007 ) ) , and jet interactions , such as the prominent examples in ngc7385 ( * ? ? ? * ( rawes 2014 ) ) and 3c390.3 . colour plotted against the @xmath15 colour . left : measured in a fixed 21-arcsec radius aperture . centre : total light from the host galaxy component , as determined from a composite srsic plus point - source model fit . right : light from the core , excluding host galaxy light to the greatest extent possible . the lines on each panel separate the region occupied by active galaxies ( upper right ) from the region occupied by passive galaxies ( lower left ) . these lines were derived from the selections of ( * ? ? ? * stern ( 2005 ) ) and ( * ? ? ? * seymour ( 2007 ) ) . only two 3crr objects lie in the active - galaxy region in total light , and only about 30% could be identified as active based on their core colours . ir surveys are therefore poor at finding radio galaxies of low or intermediate power.,title="fig:",width=167 ] colour plotted against the @xmath15 colour . left : measured in a fixed 21-arcsec radius aperture . centre : total light from the host galaxy component , as determined from a composite srsic plus point - source model fit . right : light from the core , excluding host galaxy light to the greatest extent possible . the lines on each panel separate the region occupied by active galaxies ( upper right ) from the region occupied by passive galaxies ( lower left ) . these lines were derived from the selections of ( * ? ? ? * stern ( 2005 ) ) and ( * ? ? * seymour ( 2007 ) ) . only two 3crr objects lie in the active - galaxy region in total light , and only about 30% could be identified as active based on their core colours . ir surveys are therefore poor at finding radio galaxies of low or intermediate power.,title="fig:",width=167 ] colour plotted against the @xmath15 colour . left : measured in a fixed 21-arcsec radius aperture . centre : total light from the host galaxy component , as determined from a composite srsic plus point - source model fit . right : light from the core , excluding host galaxy light to the greatest extent possible . the lines on each panel separate the region occupied by active galaxies ( upper right ) from the region occupied by passive galaxies ( lower left ) . these lines were derived from the selections of ( * ? ? ? * stern ( 2005 ) ) and ( * ? ? ? * seymour ( 2007 ) ) . only two 3crr objects lie in the active - galaxy region in total light , and only about 30% could be identified as active based on their core colours . ir surveys are therefore poor at finding radio galaxies of low or intermediate power.,title="fig:",width=167 ] the colours of the 35 3crr radio galaxies in the sample are shown in fig . [ fig : irac_colours ] . the left - hand panel shows the irac @xmath16 colours as a function of the @xmath15 colours , measured over a fixed 21-arcsec radius aperture . for only two cases do these colours suggest the presence of an active galaxy , according to the criteria of ( * ? ? ? * stern ( 2005 ) ) or ( * ? ? ? * seymour ( 2007 ) ) . that is , the average 3crr radio galaxy at low redshift appears passive . the exceptions are the two broad - line objects in the sample , 3c382 and 3c390.3 . this is as expected , since the stern and seymour selection cuts were designed to find broad - line objects . the high @xmath17 ratios evident for some of the galaxies could be interpreted as the signature of strong pah emission rather than an active nucleus . many of the higher - power sources in the sample remain in the region characteristic of pure stellar continuum , which is close to @xmath18 in this colour - colour plot . the colours of the host galaxies are shown in the centre panel of fig . [ fig : irac_colours ] . this clearly shows the passive nature of the hosts the concentration into the region of passive galaxy colours is more obvious . the most extreme departure from the locus of passive galaxies occurs for 3c293 , which is known to be gas - rich . and @xmath19 fluxes ( @xmath20 ) is shown as a function of the ratio of @xmath21 and @xmath22 fluxes ( @xmath23 ) . the position of a radio galaxy on this diagram is well - correlated with its emission - line type , except that the narrow - line source 3c326 appears anomalous , since it lies in the region occupied by low - excitation objects.,width=336 ] when the light from the host galaxies has been removed more of the 3crr galaxy nuclei show themselves to be active . the core colours are shown in the right - hand panel of fig . [ fig : irac_colours ] . the _ spitzer _ psf corresponds to a scale of more than 3 kpc at the highest redshift in this sample , and so the central stellar emission in the host galaxy is still a major component of the emission in the irac bands . this affects the irac colours , and the activity of the nuclei is only apparent for those cores which exhibit a strong warm - dust emission component . broad - line objects ( such as 3c382 ) , narrow - line objects ( such as 3c285 ) , and low - excitation objects ( such as 3c236 ) can now be found in the stern and seymour active - galaxy selection region . however , about 70% of the 3crr radio galaxies would not be selected by near - ir colour , even based on their core colours . improving the linear resolution for measurement of the ir colours of the cores would make little difference the nearby 3crr objects for which the _ spitzer _ psf corresponds to sub - kpc linear scales are not notably better separated from the locus of the passive galaxies than are the more distant objects . the line - emission characteristics of the sources are better distinguished in colours based on a combination of the mips and irac bands , as in fig . [ fig : irac_mips ] . the two broad - line emitting objects lie at the bottom of the diagram , and the low - excitation sources tend to lie in the upper left , with the narrow - line emitting cores in the lower right . 3c326 is anomalous , with ir colour more characteristic of a low - excitation source than a narrow - line - emitting source . this may indicate a misidentification of the core : 3c326 has two compact radio sources near its centre , associated with two distinct galaxies , with different line - emitting properties . the optical characteristics of the northern galaxy are more characteristic of the radio structure ( * ? ? ? * ( rawlings 1990 ) ) , but the southern galaxy matches the behaviour of other 3crr galaxies in its ir properties . the distinction between high - power and low - power radio galaxies becomes clear when the full @xmath0 spectra of the 3crr galaxies are considered ( fig . [ fig : ir_spectra ] ) . in this representative set of spectra , the low - power objects ( left panel ) show far - ir emission from dust that peaks at lower frequencies than the high - power objects ( right panel ) in most cases . residual stellar emission is also more evident at low and intermediate radio powers . the progressive change in shape of the spectra with increasing radio power strongly suggests the increasing presence of a hot dust component . the observation that there is a progression of properties also suggests that , for the majority of the 3crr sources , the cores do not change dramatically on a shorter timescale than the lifetime of the large - scale radio structure . the lifetimes of the radio sources exceed 100 myr , while the heat - loss times for dust near the cores are far shorter , @xmath24 myr . the ir spectra shown in fig . [ fig : ir_spectra ] can be fitted by the combination of a stellar continuum and dust with a range of temperatures and/or emissivitities , with the characteristic temperature of the cool component being near 60 k , and the hotter component being at about 300 k. there is no evidence for a synchrotron component in these spectra the synchrotron components in 3crr sources reported by ( * ? ? ? * cleary ( 2007 ) ) relate to quasars rather than radio galaxies . it is notable that there are no strong correlations between the core radio and ir behaviours , though we find that the core radio flux density is a slightly better predictor of the core ir flux than is the total radio flux density . there are distinct differences between the core ir spectra , particularly longward of @xmath21 , at different source powers . we believe that this indicates that the current behaviours of the cores in 3crr radio galaxies are a reasonable representation of their average behaviours over the source lifetimes . however , in detail the relationship is not strong perhaps indicating that the reservoir of material available for future activity is largely unrelated to the current nuclear activity of the sources , and that ir emission from the reservoir is a significant component of the measured fluxes . the host galaxies show no strong ir peculiarities on scales resolvable with _ spitzer _ in most cases , though some features associated with radio - source gas interactions have been noted . it is known from hst imaging that many of these objects show prominent dust lanes ( * ? ? ? * ; * ? ? ? * ; * ? ? ? * ( e.g. , de koff 1996 ; martel 1999 ; sparks 2000 ) ) , but this dust probably produces long - wavelength ir emission that contributes to the 60-k component that appears in the mips camera , and is not separated from the active cores . nevertheless , the angular resolution of _ spitzer _ is sufficient to distinguish radio hot spots , jets , and jet / cloud interactions in a significant number of cases . higher angular resolution , and improved sensitivity , would undoubtedly uncover more examples . the full team working on the 3crr low - redshift sample includes the authors of this short report and p. green , h. smith , b. wilkes , s. willner ( cfa ) , c. lawrence ( jpl ) , p. barthel ( groningen ) , e. hooper ( wisconsin ) , d , hines ( stsci ) , and i. van bemmel ( astron ) . we thank m. ashby for his advice on processing the irac images . a fuller description of this work is being prepared for publication .
we have undertaken a _ spitzer _ campaign to measure the ir structures and spectra of low - redshift 3crr radio galaxies . the results show that the @xmath0 infrared properties vary systematically with integrated source power , and so demonstrate that contemporary core activity is characteristic of the behaviour of sources over their lifetimes . ir synchrotron emission is seen from jets and hotspots in some cases . thermal emission is found from a jet / gas interaction in ngc7385 . most of the near - ir integrated colours of the low - redshift 3crr radio galaxies are similar to those of passive galaxies , so that ir colours are poor indicators of radio activity .
the glass transition from the supercooled melt to the amorphous solid still is a puzzle @xcite . while the viscosity @xmath0 increases about 15 orders of magnitude ( @xmath1 is empirically defined by @xmath2 poise ) in a narrow temperature interval , the change in the static structure factor @xmath3(@xmath4 is the scattering wavenumber ) is rather minor , see e.g. @xcite . what is then the physical basis that many different systems show a nearly universal phenomenology of relaxation \{e . g. , increase of relaxation time @xmath5 or @xmath6 $ ] described by the vogel - fulcher - tammann ( vft ) law , @xmath7 @xcite , relaxation functions obey the kohlrausch - williams - watts ( kww ) law , @xmath8 $ ] , etc . @xcite } ? the answer to this question is controversial : e.g. , there is the idea of an `` entropy catastrophe '' static phase transition underlying the glass transition \{the configurational entropy of the undercooled fluid vanishes at @xmath9@xcite , with possibly @xcite @xmath10}. mode coupling theory @xcite , on the other hand , yields a ( rounded @xcite ) dynamical transition from an ergodic fluid ( atoms easily can escape from the cages formed by their neighbors ) at a `` critical '' temperature @xmath11 to a nonergodic state ( `` structural arrest '' ) . still other concepts attribute the increase of @xmath5 upon cooling to the increase of a glass correlation length @xcite , in analogy with spin glasses @xcite . polymer melts can be held in very good metastable equilibrium in their supercooled state , and their @xmath1 s occur in an experimentally convenient range , and thus a wealth of experimental data exists . thus it is desirable to approach these systems also by theoretical modeling via computer simulation @xcite . however , due to the large size of these polymer coils ( they exhibit nontrivial structure from the scale of covalent bonds up to the gyration radius which can be 100 ) and the many decades of time scales spanned by the relaxation times of their motion simulation of polymers is very difficult , and requires to use simplified efficient models @xcite . two such models will be defined and compared in the next section , while sec . 3 describes the tests of mode coupling theory performed with these models . 4 briefly describes tests of other theories and contains also some concluding remarks . the basic idea is to eliminate both the structural details on very small scales and the associated very fast dynamics by using a coarse - graining along the chemical sequence of the chains @xcite : @xmath12 successive chemical units are mapped into one effective bond , connecting two effective monomers . these effective bonds may be defined on a lattice as well as in the continuum . in the lattice case , the effective monomers are formed from all 8 sites of an elementary cube on the lattice , and no site may belong to more than one monomer ( excluded volume interaction ) . bond lengths in this `` bond fluctuation model '' @xcite are taken in the range from 2 to @xmath13 lattice spacings , and bond vectors @xmath14 [ or permutations thereof ] are taken to be in the ground state , while all other choices of the bond vector represent excited states , which an energy cost @xmath15 also k@xmath16 ) . this bond length potential may be thought of as representing the original atomistic intrachain potentials in a coarse - grained way . since this particular choice of the potential has the effect that each bond that is in the ground state blocks the 4 sites between its adjoining effective monomers from further occupation , it has the effect that free volume in the dense system is `` wasted '' and a `` geometric frustration '' is created - at high densities not enough free volume is available that all bonds manage to get into their ground state as @xmath17 @xcite . this model can be simulated very efficiently , and for chain lengths n as small as @xmath18 properties typical of polymers ( e g. gyration radius @xmath19 ) are reproduced . a density of @xmath20 of occupied sites corresponds to a dense melt @xcite . dynamics is introduced by randomly selecting a monomer of a chain and by attempting to move it by one lattice unit in a randomly chosen direction . of course , the move is only carried out if the new position satisfies excluded volume and bond length constraints , and the possible energy change is considered with the usual metropolis criterion @xcite . it is thought that this `` random hopping '' of effective monomers in a coarse - grained sense corresponds to the random hops of chemical bonds over barriers in the torsional potential , because by such moves pieces of a chain can rotate relative to each other , and thus are responsible for the relaxation of chain configurations @xcite . if one carries out such a mapping of a realistic ( i.e. , atomistic ) chain model on the bond fluctuation model literally , one estimates that one attempted monte carlo step ( mcs ) per monomer corresponds to 10@xmath21 sec , and the lattice unit to a distance of 2 @xcite . typically , runs are carried out for systems of @xmath22 with periodic boundary conditions and @xmath23 , averaging over 16 - 160 independent `` replicas '' of the system over a time of up to 10@xmath24 mcs . thus even for the time window of this coarse - grained model we are restricted to @xmath25 ( while molecular dynamics simulations of atomistic models can deal with @xmath261nsec only ) . however , one important aspect is that one can equilibrate the system configurations also with artificial moves , which have no counterpart in the dynamics of real chains , such as the `` slithering snake '' algorithm @xcite , and thus gain several orders of magnitude in time due to faster relaxation . in this way , configurations of the well equilibrated melts are prepared at rather low temperatures , which serve as initial states for runs with the `` random hopping '' algorithm studying the dynamics of these melts . alternatively , we treat a continuum model where an effective bond is represented by a ( finitely extensible ) spring , and the effective monomers are the beads of this bead - spring model . there is a lennard - jones interaction between any pair of beads cut off at a distance @xmath27 @xcite , @xmath28 + c , \label{eq1}\ ] ] the constant c being chosen such that u@xmath29 . choosing here units of length and temperature such that @xmath30 , @xmath31 , the springs are described by @xmath32 , \qquad r_{0}=1.5,\quad k=30 . \label{eq2}\ ] ] while u@xmath33 has its preferred distance at a minimum for @xmath34 , the minimum of the potential between neighboring monomers along the chain ( given by the sum of eqs . ( [ eq1 ] ) , ( [ eq2 ] ) ) occurs for @xmath35 : although the bead - spring model is fully flexible , this incompatibility of @xmath36 with any standard crystal structure prevents crystallization of our model , and thus creates the `` frustration '' in the fluid necessary for the formation of a glass upon cooling . this off - lattice model is simulated by npt - molecular - dynamics , choosing also a chain length of @xmath18 , and altogether @xmath37 monomers in the system @xcite . note that the constant pressure ensemble @xcite is used for equilibration only - for a study of dynamic properties a clock is set to zero and runs are started at @xmath38const ( using a suitable nos - hoover thermostat @xcite ) which produce a dynamical behavior practically indistinguishable from the microcanonical ensemble @xcite . 1 shows that one obtains the structure factor @xmath39 of the amorphous polymer melt in close similarity with the experiment @xcite . if one studies the volume @xmath40 per effective monomer in a slow cooling run , starting at @xmath41 from a well - equilibrated configuration at @xmath42 and lowering @xmath43 every 500 000 md time steps by 0.02 , one finds a rather well - defined kink at @xmath44 , while a fit of the self - diffusion constant to the vogel fulcher law yields @xmath45 @xcite . thus the model does exhibit a glass transition , as expected . the bond fluctuation model also yields @xmath46 for small @xmath4 due to the very small compressibility of the polymer melt , and then a broad peak similar to the `` amorphous halo '' of fig . 1 occurs , at @xmath47 ( in units of the lattice spacing for this model ; physically this corresponds to about 1.5 @xmath48 ) @xcite . the peak position and shape also change very little with temperature . however , the second shallow and temperature - independent peak of fig . 1 at still larger q , representing intra - chain correlations , is less well reproduced @xcite . this is not surprising , since @xmath49 corresponds in real space to one lattice unit . a further disadvantage of the lattice model is that one can study the glass transition only at constant volume , while in the off - lattice model both constant volume and constant pressure studies were performed @xcite . 2 shows the intermediate incoherent dynamic structure factor @xmath50 , defined as @xmath51\}\right\rangle , \label{eq3}\ ] ] where the sum is over all @xmath52 monomers in the system ( which are at position @xmath53 at time @xmath26 ) . pronounced two - step relaxation is seen , the second step ( `` @xmath54-relaxation '' ) satisfies the time - temperature - superposition principle , while the first step ( `` beta relaxation '' ) does not @xcite . the relaxation time @xmath55 is both compatible with the vogel - fulcher law with @xmath45 , and with a power law ( fig . 3 ) but with a q - dependent exponent @xmath56 @xcite . while the consistency with a power law is evidence in favor of mct @xcite , the @xmath4-dependence of the exponent @xmath56 is not . in addition , one sees that the data deviate from the power law both for @xmath57 ( as expected , since one has left the asymptotic region ) and for @xmath58 ( which can be attributed to `` hopping processes '' by which effective monomers can escape from the cage formed by their neighbors , and requires use of the extended version of the theory @xcite that describes the rounding of the ergodic to nonergodic transition ) . thus , the observability of the idealized mct @xcite is restricted to about one decade in @xmath59 , showing the limitations of this theory when applied to polymers @xcite . in the @xmath60-regime @xmath50 can be fitted to the kww law with an exponent @xmath61 that is also weakly @xmath4-dependent ( @xmath62 for @xmath63 @xcite ) ; for a detailed analysis of this @xmath4-dependence see ref . the predictions of mct for the @xmath61 regime can also be tested in detail and compare rather favorably with mct @xcite , although the same caveat over the restricted `` temperature window '' where mct is applicable must again be made . this analysis can be replaced for several pressures and thus one can trace out a `` critical line '' @xmath64 in the ( @xmath65 ) plane separating the liquid from the ideal glass ( fig . carrying out simulations at constant pressure and at constant density that lead to the same point ( @xmath66 on the line , one finds that indeed @xmath67 holds with the same @xmath68 and the same @xmath56 for both paths in the ( @xmath65 ) plane . this underscores that the mct critical line is indeed physically significant . also the bond fluctuation model has been compared to mct , both in its idealized @xcite and extended @xcite version . while a fit of various relaxation times and of the selfdiffusion constant to the vf law is nicely compatible with the data and yields @xmath69 @xcite , the mct fits both @xcite yield @xmath70 , and power law behavior occurs over a similar temperature interval as for the continuum model . one characteristic difference between both models concerns the @xmath61-relaxation regime , however : while in the off - lattice model the first decay of the structure factor @xmath50 \{fig . 2 } is due to small - amplitude motions ( over distances of the order of 10% of the distances between effective monomers and their nearest neighbors ) , no such small scale motion is possible in the bond fluctuation model : either a monomer hops a lattice unit ) , or it can not hop at all . this distinction also shows up when we compare the time dependence of the scaled mean square displacements of monomers for both models ( fig . 5 ) . both the lattice model and the continuum model are thought to correspond essentially to the same physical system , a dense melt of short polymers , on a coarse - grained level . this implies that on mesoscopic scales ( length scales of the order of the distance between effective monomers or larger ) these different coarse - grained models should yield very similar results . comparing suitably scaled data , so that units of length and time are absorbed in the scales , this is indeed the case ( fig . 3 ) . note that the regime where mct is applicable is actually the regime of rather small displacements , up to the order of distances between monomers , from where on then a rouse - like behavior controls the dynamics . a particular advantage of the lattice model is that the slithering snake algorithm allows to obtain equilibrium properties of the model at fairly low temperatures , and also the configurational entropy @xmath71 can be computed @xcite in order to test the gibbs - di marzio theory @xcite . indeed one finds that @xmath71 decreases to about 1/3 of its high temperature value @xmath72(@xmath73 as @xmath43 approaches @xmath1 , but there the curve @xmath71 vs @xmath43 bends over and gives evidence that @xmath71 stays nonzero at lower temperatures . conversely , if one extracts the quantities that are used in the theory @xcite directly from the simulation and inserts them into the ( approximate ! ) theoretical formula @xcite , one would find an `` entropy catastrophe '' \{@xmath71 becoming negative } at @xmath74 . this unphysical result , however , is easily traced back to a severe underestimation of @xmath75 ) by the approximations of @xcite . the `` entropy catastrophe '' @xcite thus clearly is an artefact of an inaccurate approximation , one should not attribute physical significance to it . in such manner simulations can go beyond experiment for testing theories , because the input parameters of the theories can be unambiguously extracted from the simulation as well , and a comparison between theory and simulation is possible without adjustable parameters whatsoever . it is interesting to note that nevertheless the adam - gibbs equation @xcite , @xmath76 , where @xmath77 is the self - diffusion constant at very high temperatures and @xmath78 some activation energy , provides a very good description of the diffusion constants found in the simulation when one uses also the @xmath71 found in the simulation ( fig . 6 ) . at the same time , the absence of any finite size effect in fig . 6 is at odds with the idea to attribute the slowing down implied by the decrease of @xmath79 to an increasing glass correlation length @xmath80 @xcite , via @xmath81^{z}$ ] with @xmath82 a dynamic exponent @xcite : if such an hypothesis would hold , one would expect that decreasing l should decrease @xmath80 [ since @xmath83 and hence @xmath84 , and in turn @xmath79 should increase with increasing @xmath85 . fig 6 proves the absence of such finite size effects . this finding is surprising , since a growing static length can be extracted both from the pair correlation function in the melt @xcite and from surface effects near hard walls @xcite . it hence appears that there is a growing length identified in @xcite but it is not responsible for the slowing down at @xmath1 . of course , there is no contradiction with experimental results which find that @xmath1 in thin films ( or in pores , respectively ) changes when the linear dimension of the film ( or pore respectively ) is varied : depending on the boundary conditions at the surface , the local mobility of monomers in the surface region changes , and this effect is the more pronounced on the freezing the smaller the linear dimension . no such surface effect is present with the periodic boundary conditions in fig . 6 , of course . in experiments on the glass transition in confined systems , finite size effects and surface effects can never be clearly separated , unlike simulations where one can show that there is no finite size effect ( fig . 6 ) at least in the temperature region studied here , although there do exist surface effects @xcite . returning to mct , we emphasize that idealized mct does provide a good description of a large number of simulation data , but only over a rather restricted range of temperatures ( about one decade in @xmath59 ) and corresponding times ( or viscosities , typically the range @xmath86 poise if @xmath87 poise : the huge range from 10@xmath88 poise is outside the scope of the theory ) . correspondingly , only a small intermediate regime of small monomer displacements in fig . 5 is described - neither the initial increase that depends on microscopic properties of the model , nor the regime of hopping processes that lead to a rouse - like relaxation of coil configurations ( before ultimately ordinary diffusion sets in ) are part of the theory . since there occur smooth crossovers between the various regimes , a reliable assessment of the validity of mct for polymer melts near @xmath1 is difficult . a more complete theory ( that unifies e.g. mct and the rouse model ) would be very desirable . we are grateful to s. bhmer , k. okun , v. tries and m. wolfgardt , who have contributed to earlier stages of this work , for their fruitful collaboration . we acknowledge the financial support from the deutsche forschungsgemeinschaft ( dfg ) , grant no sfb262/d2 , and we thank the hchstleistungsrechenzentrum jlich ( hlrz ) and the regionales hochschulrechenzentrum kaiserslautern ( rhrk ) for generous grants of computer time . c. bennemann , w. paul , j. baschnagel , and k. binder _ investigating the influence of different thermodynamic paths on the structural relaxation via glass forming polymer melt _ , submitted to j. phys . : condens . matter .
two coarse - grained models for polymer chains in dense glass - forming polymer melts are studied by computer simulation : the bond fluctuation model on a simple cubic lattice where a bond - length potential favors long bonds is treated by dynamic monte carlo methods , and a bead - spring model in the continuum with a lennard - jones potential between the beads is treated by molecular dynamics . while the dynamics of both models differ for short length scales and associated time scales , on mesoscopic spatial and temporal scales both models behave similarly . in particular , the mode coupling theory of the glass transition can be used to interpret the slowing down of the undercooled polymer melt . for the off - lattice model , the approach to the critical point of mode coupling is both studied for constant pressure and for constant volume . the lattice model allows a test of the gibbs - di marzio entropy theory of the glass transition , and our finding is that although the entropy does decrease significantly , there is no `` entropy catastrophe '' where the fluid entropy would turn negative . finally , an outlook on confinement effects on the glass transition in thin film geometry is given . contribution to the ii workshop on non equilibrium phenomena in supercooled fluids , glasses and amorphous materials , pisa 1998 , submitted to _ j. phys . : condens . matter _ on sept . 27 , 1998
in this section we summarize the experimental results from the heavy flavour working group @xcite . the presentations covered a wide range of topics , from charm , beauty and charmonium production in @xmath8 and @xmath9 collisions , heavy ion results on charm suppression , spectroscopy and rare decays , over oscillations in the @xmath2 and @xmath3 systems to the outlook for heavy flavour physics at future experiments at the lhc and the ilc . charm production in @xmath8 collisions has been studied extensively over the last years at hera , and a wealth of data exist in photoproduction ( where the exchanged photon is quasi - real , with a virtuality @xmath10 ) and deep - inelastic scattering ( dis ) . the hera collider experiments , h1 and zeus , have presented new preliminary results on charm production from the hera - ii running period @xcite . the zeus collaboration reported on two charm measurements in dis with hera - ii data ( fig . [ fig : zeus - f2c ] ) : one analysis @xcite uses @xmath11 mesons to identify charm production , and utilizes @xmath12 of new data to achieve improved statistical accuracy compared to previous analyses . the second analysis @xcite uses @xmath13 mesons instead and is one of the first measurements to utilize the new silicon strip micro vertex detector ( mvd ) of zeus . based on @xmath14 , this analysis achieves similar accuracies as previous zeus measurements of the inclusive charm cross section in dis . the zeus collaboration has recently also finalized two analyses of charm production from older hera - i data @xcite . one analysis @xcite covers the region of very low momentum transfer ( @xmath15 ) at the transition between photoproduction and dis . the data provide a good test of perturbative qcd calculations , which are available in nlo , and are well described by massive calculations in the fixed flavour number scheme , as implemented in the hvqdis program @xcite . in the second analysis @xcite , the pseudoscalar states @xmath16 , @xmath17 , and @xmath18 are reconstructed , rather than the vector state @xmath19 , which allows a measurement of the fragmentation fractions , which turn out to be compatible with those measured in @xmath20 annihilation and in photoproduction @xcite . the inclusive charm cross section derived from this measurement is consistent with previous results . also a new , preliminary measurement of charm fragmentation was presented by the zeus collaboration @xcite , which shows broad agreement with other measurements from h1 @xcite , and also with measurements from @xmath20 experiments . measurements from zeus . ] another new measurement of @xmath11 production in dis , based on @xmath21 of hera - ii data , was presented by the h1 collaboration @xcite . here , differential and double differential visible cross sections for @xmath11 production were measured and compared to the qcd calculations . overall , these quantities are well described by the nlo predictions of hvqdis ; in fact , the data uncertainties are in many cases smaller than the theory uncertainties from the variation of the charm mass and the variation of the renormalization and factorization scales . however , in some quantities , most notably the @xmath22 distribution , even in spite of the relatively large theory uncertainties the nlo predictions deviate significantly from the trend observed in data . this confirms , with higher statistics , the observation made in a recently published analysis from the h1 collaboration @xcite . however , that analysis goes one step further and investigates @xmath11 production in dis in conjunction with jets . as heavy flavour production is dominated by boson gluon fusion , a two - jet structure is expected for most of the events . in the dis analysis , one jet in addition to the @xmath11 meson is required . again , hvqdis describes the data satisfactorily , with the notable exception of the azimuthal angle difference @xmath23 between the @xmath11 and the jet , a quantity which would be almost exactly equal to @xmath24 in leading order , where the two charm jets must be back - to - back ; hard gluon emission , which enters only in nlo , leads to an enhancement at lower values of @xmath23 , which is underestimated by hvqdis , whereas monte carlo models such as cascade @xcite , which include parton showers as approximation for higher - order gluon radiation , work significantly better . in another recent publication @xcite , the h1 collaboration has investigated charm photoproduction with two jets , with very similar findings . the results show that charm measurements have now reached a level of accuracy where more precise qcd predictions would be desirable . on one hand a full nlo monte carlo with matched parton showers is needed , as opposed to hvqdis , which produces only parton four - vectors and has been augmented with relatively simple independent fragmentation routines to make the extraction of visible cross sections possible . a second way to reduce theory uncertainties would be to reduce the variation of input parameters such as the charm mass or fragmentation parameters by detailed comparisons between mc predictions and data of ( double ) differential visible cross sections in terms of relevant quantities such as @xmath25 and @xmath22 . on the other hand , nnlo calculations would provide the most reliable way to reduce theory uncertainties , in particular for inclusive quantities such as @xmath26 . beauty production at hera and the tevatron has generated a lot of interest in recent years , not least because at both colliders inclusive cross sections for beauty production were observed that are considerably higher than the expectation from perturbative qcd calculations in next - to - leading order . achim geiser gave an overview over the existing experimental results from hera , with additional glimpses at tevatron and ua1 data @xcite . he discussed in detail the issue of scale dependence of qcd calculations , where he presented a survey of qcd calculations in different orders of @xmath27 for a number of processes and argued that a good choice for the factorisation and renormalisation scales @xmath28 would be a scale at which higher order calculations give a result that coincides with the calculations at lower order . he observed that this is achieved in most of the cases at a scale between @xmath29 and @xmath30 , where @xmath31 is the `` natural '' choice of scale such as @xmath32 , @xmath33 , or @xmath34 . based on this observation he argued that the factorisation and renormalisation scales should be varied in the range @xmath35 , with a choice of @xmath36 for the central value , rather than using a variation of @xmath37 , as is customary in most hera publications . he also pointed out that this prescription has already been adopted silently by the tevatron experiments in recent preliminary results . this proposal was received well by the audience , as the discussion after the presentation showed . new , preliminary results on beauty production were presented by the d0 and cdf collaborations @xcite . d0 has performed a new measurement of muon tagged jet cross sections based on @xmath38 of data that extends up to @xmath39 . again , the data lie above the theory prediction . in a new measurement of inclusive b cross sections , the cdf collaboration utilizes the decays @xmath40 in addition to the previously used channel and @xmath41 , with consistent results . these measurements are also consistent with a fixed order calculation at next to leading log . the first , preliminary hera - ii result on beauty production , based on @xmath42 of data , was presented by the zeus collaboration @xcite . the results are somewhat higher , but still consistent with previous measurements from h1 @xcite . the inclusive production of charmonium and bottomium states in @xmath8 and @xmath43 collisions remains an intersting testing ground for perturbative qcd , as was shown in a review by katja krger @xcite . the working horse for these investigations are @xmath44 mesons , where the largest data sets are available . while leading order calculations in the colour singlet model @xcite fail to describe the production rates of @xmath44 mesons at the tevatron @xcite , nrqcd models @xcite models have been successfully applied there . in these models the production of @xmath45 states with different colour and spin / parity configuration is calculated perturbatively , and the transition to bound states is described by non - perturbative long distance matrix elements ( ldme ) , which have to be determined from data . this allows the production of @xmath44 mesons by gluon splitting into a @xmath45 pair , followed by a transition to a @xmath44 meson , which is predicted to be transversely polarized . the observation of such a polarization is therefore considered the `` smoking gun '' for ( large ) colour octet ( co ) contributions . however , new data from cdf that were presented at this conference @xcite show a clear evidence for longitudinal @xmath44 polarization , which is in contrast to nrqcd expectations . new data on @xmath46 polarisation are still not precise enough for firm conclusions . a second test of nrqcd predictions is the measurement of the production rates of @xmath47 versus @xmath48 states , which in nrqcd models is expected to follow the spin counting prediction of @xmath49 . new measurements from cdf @xcite give a significantly smaller result for the ratio of prompt @xmath50 production : @xmath51 after the initial success of the nrqcd model indications for the necessity of large co contributions where also searched for in inelastic @xmath44 production at hera . the h1 collaboration has presented new , preliminary data from hera - ii on inelastic @xmath44 production in deep inelastic @xmath8 scattering @xcite . production rates were measured double differentially in the transverse momentum @xmath52 and the momentum fraction @xmath53 of the @xmath44 meson , and compared to monte carlo predictions by programs which implement the colour singlet model in leading order . these models describe the shape of the measurements quite well , whereas the co contributions in nrqcd models tend to have different @xmath52 and @xmath53 shapes than the cs contributions . therefore , the new h1 data do not show any clear need for large co contributions to inelastic @xmath44 production . in the discussion of the results the need was stressed for nlo calculations of this process , in the cs as well as in nrqcd models . nlo calculations are currently only available in the cs model for photoproduction of @xmath44 mesons @xcite , where they describe the data from hera @xcite quite well . a different issue in charmonium production is adressed by the new data from hera - b @xcite : hera - b has collected large samples of @xmath44 and @xmath46 mesons decaying to @xmath54 and @xmath20 pairs , which made it possible to investigate the dependence of charmonium production on the atomic weight @xmath55 in a new range of feynman-@xmath56 , extending the range covered by the experiments down to @xmath57 , which is a region where theoretical models make widely different predictions . the new results from heavy ion experiments at rhic have been reviewed in this conference by william zajc @xcite . therefore we only briefly highlight some of the new results presented in the heavy flavour session . new results on heavy quark production in au@xmath58au collisions were presented by the phenix @xcite and the star @xcite collaborations . due to their larger mass and the dead cone effect , charm quarks are expected to lose energy at a smaller rate than light quarks . the observation of supression of electrons from charm decays in central au@xmath58au collisions @xcite therefore comes as a surprise . it appears that charm quarks participate in the flow of the opaque hadronic medium just as much as light quarks , which means that they thermalize more quickly than expected in many theoretical models . the phenix collaboration has also presented new measurements of @xmath44 suppression in au@xmath58au collisions @xcite . these measurements , which are performed in two rapidity ranges @xmath59 and @xmath60 , show a significant suppression of @xmath44 production in au@xmath58au collisions , which is stronger at large rapidities than at central rapidity values . the explanation of this rapidity dependence is a real challenge to theoretical models . the large data sets from the b factory experiments babar and belle , from cleo - c and from the tevatron experiments cdf and d0 have lead to a renewed interest in the spectroscopy of charm and bottom hadrons with beautyful results , and have opened the possibility to identify extremely rare decays . in the charmonium system , finally all expected charmonium states below the d - meson threshold have now been firmly identified . one of last two missing states , the @xmath61 , has now seen by babar , cleo - c , and belle @xcite , at an average mass of @xmath62 . this allows a comparison of the hyper fine splitting of the @xmath63 and @xmath64 states in the charmonium system , which are @xmath65 and @xmath66 ; the large difference of these values poses a challenge to theory @xcite . the long elusive @xmath67 state has observed by cleo - c @xcite in the decay chain @xmath68 , @xmath69 , with a mass of @xmath70 , which corresponds to a hyper fine splitting between the @xmath67 and the center of gravity of the @xmath71 states of @xmath72 , consistent with the expected value of zero . meanwhile , cleo - c has increased its @xmath73 sample eightfold , which yield the promise of further , improved results . in addition to these expected charmonium states , recent years have seen the discovery of several unexpected charmonium - like resonances : new results were obtained on the @xmath74 , which is already considered firmly established by the pdg @xcite , and on the @xmath75 and @xmath76 . results were also presented on two additional states , the @xmath77 and the @xmath78 . for the @xmath74 , one explanation that has been put forward is the interpretation as a @xmath79 molecule . cleo has performed a new measurement of the @xmath16 mass @xcite : @xmath80 @xcite . the total uncertainty of this measurement of @xmath81 is a factor of 2.2 better than the uncertainty of @xmath82 of the 2006 world average @xcite . combinig this result with the pdg06 value of the @xmath74 mass of @xmath83 results in a very small binding energy of @xmath84 for a @xmath79 molecule @xcite . meanwhile , belle and babar have also found indications for the @xmath74 in the decays @xmath85 ( belle ) and @xmath86 ( babar ) @xcite . in these channels , the observed mass values for the @xmath74 of @xmath87 ( belle ) @xcite and @xmath88 ( babar ) are @xmath89 higher than the current world average . belle also concludes that the quantum numbers @xmath90 are favoured if the observed enhancement is indeed the @xmath74 . all in all , the interpretation of this state remains unclear . new data were also presented on the @xmath76 @xcite , which was first observed in initial state radiation events at babar . babar sets a limit @xcite of @xmath91 , which is a further indication that the @xmath76 is not a conventional charmonium state . cleo has also confirmed the @xmath76 @xcite , and finds @xcite its mass to be @xmath92 , in poor agreement with the original babar measurement @xcite of @xmath93 . the latest belle @xcite result @xmath94 is consistent with the cleo measurement . the discovery of new , unexpected charmonium - like states has also triggered new investigations of @xmath95 at belle and cleo @xcite . these scans show a marked dip around the @xmath76 . belle , babar and cleo have also looked @xcite into the more exclusive channels @xmath96 . a weak signal for the @xmath76 is seen in the @xmath97 channel , no signal in the @xmath98 channel , and a dip , similar to the one observed in the inclusive @xmath99 measurement , is observed in the @xmath100 channel around the @xmath76 resonance . meanwhile , babar observes yet another state @xcite in initial state production of @xmath101 at a mass of @xmath102 and a width of @xmath103 , which is incompatible with the @xmath76 or other known states such as the @xmath104 . three more states observed by belle @xcite , termed @xmath77 , @xmath75 , and @xmath78 , may have an interpretation as conventional charmonium states , namely the @xmath105 $ ] , the @xmath106 $ ] and the @xmath107 $ ] . in the sector of charmed , strange mesons new measurements were presented by babar and belle @xcite . in addition to new measurements of the properties of the @xmath108 and @xmath109 , another new state , termed @xmath110 , has been identified by babar @xcite with a mass @xmath111 and a spin parity assignment @xmath112 . in addition , a hint for another state with @xmath113 has also been observed . furthermore , belle , in a dalitz analysis of the decay @xmath114 , sees indications for a state @xmath115 with @xmath116 , with quantum numbers @xmath117 favoured . the theoretical interpretation of these states is not yet clear . while the @xmath108 and @xmath109 states can be explained as @xmath118 and @xmath119 p - wave @xmath120 states , their masses are substantially lower than expected from potential models . the interpretation of the other states is still less clear . coming to charmed baryons , babar and belle have reported discoveries of several new states @xcite . both observe the new state @xmath121 in the channels @xmath122 ( babar ) and @xmath123 ( belle ) @xcite . the belle discovery of the new charmed , strange baryons @xmath124 and @xmath125 has been confirmed by babar @xcite , and belle sees also some evidence for the isospin partners @xmath126 and @xmath127 . babar has also reported the first observation of the @xmath128 , an excited @xmath129 state @xcite . in the b meson sector , the progress comes from the tevatron experiments cdf and d0 @xcite . the @xmath130 has already been observed by both experiments , now cdf reports the first direct observation of the @xmath130 in the exclusive decay channel @xmath131 , which allows a very precise mass measurement with the preliminary result of @xmath132 , while the uncertainty on the old world average was @xmath133 . the @xmath134 is last spin singlet ground state in the bottomium sector that has not been unambiguously discovered yet . cdf has performed a new search for the decay @xmath135 , without indications for a signal . both , d0 and cdf , report first direct observations on orbitally excited b mesons with @xmath136 by looking for decays @xmath137 . both see clear evidence for @xmath138 and @xmath139 states . with a similar analysis of the channels @xmath140 both experiments observe also the @xmath141 state , in addition cdf reports evidence for the @xmath142 state . in the bottom baryon sector , where up to now the @xmath143 is the only well established particle , cdf performed a blind analysis of the decay channel @xmath144 ; after unblinding , four resonances were found with significances greater than @xmath145 , which constitute the first direct observation of @xmath146 baryons . the resonances are identified as the @xmath147 and the @xmath148 . in the field of rare decays , the large data sets of altogether more than 1 billion @xmath149 events obtained at the b factory experiments allow more and more precise measurements of @xmath150 decay modes @xcite , which are interesting because in the standard model they are forbidden at tree level and proceed through loop diagrams . in extensions of the standard model , additional particles contribute to the loops , which may lead to observable deviations of the transition rates from the expected sm values . while the @xmath151 decay channels are investigated with higher and higher precision , the large statistics now allows even the measurement of @xmath152 transitions , which were first observed by belle in 2005 @xcite . the latest compilation from the heavy flavour averaging group @xcite contains 20 measured @xmath153 decay channels , with branching fractions as small as @xmath154 for @xmath155 and precisions down to @xmath156 for @xmath157 . the 2006 average for @xmath151 decays @xcite is now @xmath158 , which is on the high side of recent nnlo predictions @xcite . the latest measurements from babar and belle of @xmath159 @xcite are particularly impressive , measuring branching fractions around @xmath160 , in some cases with more than @xmath145 significance . from a combination of the @xmath161 measurements from belle and babar , a constraint of @xmath162 has been derived . a very difficult , but interesting decay channel is @xmath163 @xcite , which has been observed by belle at @xmath164 significance with a branching ratio of @xmath165 , while babar measures @xmath166 , or , translated into a @xmath167 cl limit , @xmath168 . the combination of both results yields @xmath169 , corresponding to a @xmath89 evidence . this result can be used to derive limits on the @xmath170 mass in susy models . after the observation of @xmath171 oscillations , the search for the decays @xmath172 might be considered the next race in b physics . the standard model expectations for these decays are extremely small , and probably out of reach for current experiments : @xmath173 and @xmath174 . again , these decays can only proceed through loop diagrams in the sm , and contributions from new particles may increase the rate by orders of magnitude @xcite . both tevatron experiments have searched for the @xmath175 decay and have reported new , preliminary results based on new run - ii data . the limits are @xmath176 from cdf ( d0 ) ; these limits correspond to branching ratios that are 29 ( 27 ) times larger than the sm expectation . cdf has also reported a preliminary limit for @xmath177 decays : @xmath178 , which is 230 times greater than the sm expectation . while oscillations have long been established and thoroughly investigated in the @xmath179 and @xmath180 systems , until recently the frequency of the @xmath2 oscillations had not been measured , and no firm signal for mixing in the @xmath3 system had been observed . during the last year , both these gaps in our knowledge of neutral meson mixing have been filled . the @xmath171 oscillation frequency has now been measured by the tevatron experiments @xcite . after the first report on a double sided limit for the @xmath171 oscillation frequency of @xmath181 at @xmath167 cl from the d0 collaboration @xcite , cdf has for the first time observed a nonzero oscillation amplitude with more than @xmath182 significance @xcite , at a frequency @xmath183 , consistent with the d0 result . meanwhile , the d0 collaboration has gone several steps further . based on a sample of @xmath184 of decays @xmath185 , d0 has made a new measurement @xcite of the difference between the lifetimes of the long and short lived @xmath171 eigenstates . both states can decay to the @xmath186 final state , and by fitting the decay angle distributions their respective contributions to the sample at different eigentimes can be determined , with the result @xmath187 . for the first time , the d0 collaboration has used this data also to extract the cp violating phase @xmath188 , which is the relative phase of the off - diagonal elements of the mass and decay matrices in the @xmath189 basis , from this data : @xmath190 . the sm prediction for @xmath188 is very small , namely @xmath191 @xcite . d0 has also reported on a new measurement @xcite of the branching ratio @xmath192 , which is lower than the only pre - existing measurement from aleph . this branching ratio is linked theoretically to the width difference ( wrong - sign ws ) decays from babar ] @xmath193 between the cp - even and odd @xmath171 eigenstates . the resulting constraint is @xmath194 . while the difficulty in @xmath195 mixing lies in the fact that the mixing is almost perfect because the oscillations occur much faster than the decay , the situation is reversed in the @xmath196 system . here , the standard model prediction for the mixing parameters @xmath197 and @xmath198 are very small @xcite , of the order @xmath199 , which means that the @xmath16 decays much faster than one oscillation period lasts . again , additional particles in the loop may increase the mixing , which could indicate new physics . in particular , new physics processes such as flavour - changing neutral currents , susy particles etc . would increase @xmath200 ( i.e. , the oscillation frequency ) , but not @xmath201 ( the lifetime difference ) . until march of this year , the no - mixing case had been disfavoured with a significance of @xmath202 by belle ( babar ) analyses @xcite . then both collaborations published two papers titled `` evidence for @xmath3 mixing '' side by side in physics review letters @xcite , both reporting more than @xmath182 evidence for @xmath3 oscillations . the belle analysis is based on @xmath203 of data and measures the difference in apparent lifetime for the cp even decays @xmath204 to the lifetime of the decay @xmath205 . the data show indications for mixing with @xmath206 significance . in addition , the belle collaboration has performed a dalitz analysis of the decay @xmath207 , which provides the most sensitive result on @xmath200 to date : @xmath208 and disfavours the no - mixing solution with @xmath209 significance @xcite . the babar analysis uses @xmath210 of data and analyses the decays @xmath211 . the dominant decay mode is the right - sign ( rs ) decay mode to @xmath212 , while the wrong - sign ( ws ) decay mode to @xmath213 may occur as doubly cabibbo - suppressed ( dcs ) decay or through mixing . these two mechanisms are separated by the analysis of the time dependence of the decay , see fig [ fig : babar - d0-ws ] . from this data , babar can exclude the no - mixing hypothesis with @xmath214 . looking ahead into the future , first beams at the lhc are now expected in early 2008 . heavy flavour physics will be of interest at the experiments atlas , cms @xcite and lhcb @xcite . in atlas and cms , the expected rates for beauty production are high : at a luminosity of @xmath215 , around @xmath216 @xmath217 pairs will be produced per second , which should allow high precision measurements despite the difficult environment . the big challenge here is to set up a reasonably efficient trigger , because bottom quarks are predominantly produced at relatively low transverse momenta , while the detectors are optimized for high-@xmath218 discovery physics . in the initial , `` low '' luminosity phase with expected luminosities around @xmath215 , the atlas experiment envisages to use single particle triggers at level 1 , which will be refined by searches for more complex signatures in the high level trigger ( hlt ) @xcite . channels that have been studied include e.g. @xmath219 with subsequent decays @xmath220 , states involving a electromagnetic signature such as @xmath221 , and final states with two muons . in the high luminosity mode at @xmath222 , level 1 will predominantly use the dimuon channel to identify @xmath44 or @xmath223 decays to @xmath224 . cms will also use muon triggers at level 1 , and plans to identify decay vertices from heavy flavour decays in the high level trigger . a number of topics have been studied by both collaborations , such as the measurement of inclusive beauty cross sections ( cms ) , the measurement of @xmath225 in the golden mode @xmath226 ( atlas ) , analysis of @xmath185 decays ( atlas and cms ) , and searches for rare decays such as @xmath227 , @xmath228 or @xmath229 ( atlas ) . the atlas collaboration has also studied the prospects to measure @xmath171 oscillations . they conclude that they could measure @xmath230 with @xmath145 significance at the current cdf value with the data from one year , i.e. @xmath231 . cms has also studied the prospects to measure the @xmath232 mass from decays to @xmath233 ; the expected resolution from @xmath234 of data is @xmath235 statistical and @xmath236 systematical uncertainty , which can be compared to the preliminary cdf result reported in this conference @xcite , based on @xmath184 , with statistical and systematic errors of @xmath237 and @xmath238 , respectively . the lhcb experiment @xcite , is an experiment dedicated to the study of b physics at lhc . the detector is a single arm spectrometer , optimized to detect b hadrons at pseudo rapidities @xmath239 , and will run at an interaction zone with a relatively low luminosity to reduce backgrounds , where one year of running at nominal luminosity will provide @xmath240 of data . the lhcb collaboration has studied the prospects to measure the angles of the unitarity triangle ; for @xmath241 they hope to achieve a statistical precision of @xmath242 from one year of data taking in the golden channel @xmath226 . for the least well determined angle @xmath243 , the most promising method seems to be the analysis of decays @xmath244 with @xmath16 decays to @xmath245 and @xmath246 , with a comparison of the rates of cabibbo favoured and doubly cabibbo suppressed decays . the estimated sensitivities are in the range @xmath247 , depending on the actual value of @xmath243 . an interesting benchmark is provided by the sensitivity of @xmath248 , which has been studied by atlas , cms @xcite , and lhcb @xcite . recall that currently cdf and d0 have reported upper limits of @xmath249 for the branching ratio @xcite based on samples of @xmath240 ( d0 ) and @xmath250 ( cdf ) ; these limits lie about 30 times above the sm expectation of @xmath251 . a recent atlas study concludes that with an integrated luminosity of @xmath252 , corresponding to about three year s running in the low luminosity mode , an upper limit of @xmath253 at @xmath167 cl is achieveable , which is about a factor of two above the sm expectation . cms concludes that it could provide a limit of @xmath254 ( four times the sm expectation ) with @xmath231 , i.e. one year s worth of data . lhcb on the other hand , which will operate at lower luminosities , claims a sensitivity that would allow a @xmath182 observation of the decay @xmath255 at the standard model branching ratio of @xmath251 with @xmath240 , corresponding to one year s worth of data @xcite at nominal luminosity . looking even farther into the future , tim greenshaw reported on the applications of heavy flavour at the planned innternational linear collider ( ilc ) @xcite , an @xmath20 collider with a centre - of - mass energy between @xmath256 and @xmath257 . the detector designs for this machine forsee a very good flavour separation power , which calls for vertex detectors with spectacular performance . several groups are currently investigating various technologies that could fulfill the requirements for the inner pixel detector of an ilc detector . one important application of the heavy flavour identification capabilities of such a detector would be the measurement of the various branching fractions of a higgs boson . even if only one neutral higgs boson were found at the lhc , this would allow to check wether these fractions are consistent with the predictions from the standard model , where only one higgs doublet is assumed and therefore all branching fractions depend only on the mass of the single physical higgs boson , or whether for instance up- and down - type fermions have different couplings to the higgs , as expected for instance in supersymmetric models , which have at least two higgs doublets . heavy flavour production at hera and tevatron remains an interesting field of research , both from the experimental and the theory side . the experimental results become more and more precise , which makes it interesting to include them in the determination of parton densities , provided that the theoretical obstacles can be overcome and uncertainties arising from charm mass and fragmentation functions are treated consistently between experiment and theory . the beauty production data have taught us that the calculation of production cross sections do not always become more accurate for heavier quarks if there is sufficient phase space for qcd dynamics , which is something to keep in mind for lhc , where also the top will fall into this category . during the last years we have seen a veritable renaissance of hadron spectroscopy . as the particle data book fills up , we see more and more results that indicate how incomplete our understanding of hadron structure still is , as illustrated by unexpected differences in hyperfine splittings in the charmonium sector @xcite , by unexpected states and masses in the @xmath258 system @xcite , and by the appearance of charmonium - like resonances for which we have no ready explanation @xcite . while resonances that do not fit into the conventional @xmath259 picture have been known for a long time in the light meson sector , we now learn that also in the heavy quark sector there is more than is written in our philosophy . on the other hand , the investigation of rare decays such as @xmath260 or @xmath261 and other rare processes such as @xmath3 mixing , where we wait for results that would point to physics beyond the standard model , has once again failed to turn up anything unexpected . in this section , we summarize the six theory contributions to the heavy - flavor working group , emphasizing computations that have been performed during the years 2006 and 2007 and have therefore not yet been presented at previous dis workshops @xcite . we start with new perturbative results for inclusive final states , i.e. heavy - quark structure functions and their relation to parton densities , and then move on to less inclusive final states , in particular quarkonium production and decay and heavy - quark spectroscopy . for the latter we emphasize new results from lattice qcd and qcd sum rules . note that a series of new calculations on open heavy - quark production at various colliders in the general - mass variable - flavor number scheme ( gm - vfns ) has been presented at this workshop in an introductory plenary talk by g. kramer @xcite . the heavy - quark coefficient functions for deep - inelastic lepton - hadron scattering ( dis ) in the kinematic regime @xmath262 have been calculated more than ten years ago in next - to - leading order ( nlo ) of qcd by m. buza et al . , using operator product expansion techniques @xcite . here @xmath263 and @xmath264 stand for the masses squared of the virtual photon and heavy quark respectively . the analytical results had been expressed in terms of 48 independent functions and had been used to check earlier , general calculations , which were , however , only accessible via large computer programs . j. blmlein has now presented a re - calculation of the @xmath265 massive operator matrix elements for the twist-2 operators , which contribute to the heavy flavor wilson coefficients in unpolarized dis in the region @xmath266 , using light - cone expansion techniques and confirming the above - cited calculation @xcite . the application of the integration - by - parts method and harmonic sums in mellin space allowed for a significant compactification of the results , which can now be expressed in terms of the basis @xmath267 and @xmath268 , i.e.of only two independent functions . while the proton is just a simple @xmath269 fock state in the quark model , the possibility of an intrinsic - charm , i.e. a @xmath270 , component has repeatedly been put forward in the context of light - cone @xcite or meson - cloud models @xcite . tung et al . have performed global fits of parton density functions ( pdfs ) , assuming that the charm - density is not only generated radiatively at @xmath271 and then evolved to @xmath272 , but allowing for the possibility of light - cone , meson - cloud , or sea - quark like intrinsic charm density @xcite . the quality of each fit is measured by a global @xmath273 , shown in fig . [ hfl:1 ] as a function of the momentum fraction @xmath274 carried by the charm quark at the starting scale @xmath275 gev . for @xmath276 , the quality of the fit varies very little , i.e. the global analysis of hard - scattering data gev for the light - cone ( solid curve ) , meson - cloud ( dashed curve ) , and sea - like model ( dotted curve ) . ] provides no evidence either for or against intrinsic charm . above this point , all three curves in fig.[hfl:1 ] rise steeply with @xmath277 , so that global fits do place useful upper bounds on intrinsic charm . there is no data set that is particularly sensitive to intrinsic charm , but the global qcd analysis rules out the possibility of an intrinsic charm component much larger than @xmath242 in momentum fraction . a variation of the charm quark mass @xmath278 shows that the data prefer lower masses around 1.3 gev with respect to higher masses of 1.5 gev . the difference in @xmath273 is in this case almost entirely due to the charm contribution to the proton structure function @xmath279 , which has been precisely measured at hera . turning to less inclusive quantities , the production and decay of heavy quark - antiquark bound states ( quarkonia @xmath280 ) is still far from understood . while the effective field theory of non - relativistic qcd ( nrqcd ) has long been believed to be phenomenologically successful and is still believed to be theoretically more consistent than the color - singlet model ( csm ) , recent tevatron run ii data on the polarization of large-@xmath281 @xmath282- and @xmath283-mesons do not support the idea that their production is dominated by color - octet fragmentation processes as predicted in nrqcd . since higher - order qcd corrections are largely unknown ( except for color - singlet photoproduction @xcite , color - singlet and color - octet production in photon - photon collisions @xcite and a very recent calculation for direct color - singlet hadroproduction @xcite ) , other theoretical frameworks continue to be investigated . one example is the @xmath284-factorization formalism , where the production cross section @xmath285 is computed by a double convolution of unintegrated parton densities @xmath286 and partonic cross section @xmath287 over longitudinal momentum fractions @xmath288 and intrinsic transverse momenta @xmath289 of the reggeized partons @xmath99 in the ( anti-)proton @xmath290 ( @xmath291 ) . this method has been applied by kniehl et al . to charmonium production at different colliders @xcite and now , as v.a . saleev reported , also to bottomonium hadroproduction at the tevatron @xcite . the long - distance operator matrix elements ( omes ) of nrqcd have been refitted to the @xmath281-spectra of prompt @xmath283 mesons at the tevatron . since the intrinsic @xmath284 leads to a harder @xmath281-spectrum already for the color - singlet contributions , the color - octet omes turn out to be considerably smaller and in many cases even consistent with zero , as their values are not sufficiently constrained by the data . in addition , the results depend strongly on the assumed unintegrated parton densities , which are poorly known and only constrained to agree with the integrated ones , @xmath292 as can be seen from tab . [ hfl:2 ] , the quality of the fit , measured .[hfl:2]omes of the @xmath293 and @xmath294 mesons from fits to cdf data from runs i and ii in the regge - kinematics approach using unintegrated gluon distributions by j. blmlein ( jb ) , jung and salam ( js ) , and kimber , martin and ryskin ( kmr ) . [ cols="^,^,^,^",options="header " , ] theoretical uncertainties with the chosen masses and scales and with respect to possible color - octet contributions has not been performed . the calculated @xmath44-decay branching ratios to the @xmath295-wave mesons @xmath296 and @xmath297 fit the data well , while that of @xmath298 ( if treated as an @xmath299 meson ) is predicted to be @xmath300 , which implies that @xmath301 can not be the @xmath299 or @xmath302 meson . decays of @xmath295-wave charmonia @xmath303 ( @xmath304 ) have also been studied . the branching ratio of @xmath305 is predicted to be @xmath306 , which may be tested by cleo - c and besiii with future experiments . a similar calculation has been performed for the radiative decays of bottomonia into charmonia and light mesons , taking into account in addition the sometimes significant qed contributions @xcite . while the results for radiative decays into truly non - relativistic charmonia are likely to be more reliable than those for radiative decays into ( in reality relativistic ) light mesons , these decays remain to be observed . on the other hand , the calculated branching ratios for @xmath307 ( @xmath308 ) and @xmath309 ( @xmath310 ) are only barely consistent with recent cleo data [ @xmath311 and @xmath312 , suggesting that the theoretical approach may still need improvement . finally , j.m . richard reviewed the current status of light and heavy multiquark spectroscopy @xcite . while experimental evidence for the @xmath313 pentaquark @xmath314 seems to vanish , a single - charm baryon with a mass of 2940 mev and a width of approximately 17 mev has recently been confirmed by babar , and a double - charm baryon with a mass of 3520 mev may have been observed by the selex collaboration . potential models have since long predicted the existence of baryons with at least one heavy quark , assuming them to be diquark - quark systems that can be described with hyperspherical coordinates . in the meson sector , the state @xmath77 may be a @xmath315 candidate for a @xmath316 hybrid state , to be described in the born - oppenheimer approximation with classical constituent gluons . a second recently confirmed state is the @xmath74 , which qcd sum rules predict to be a @xmath317 @xmath318 bound state , even if the theoretical mass found is slightly larger ( @xmath319 mev ) @xcite . we thank all the speakers of the heavy - flavor working group for their contributions and the local committee for the perfect organization of the dis 2007 workshop . s. chekanov _ et al . _ [ zeus collab . ] , phys . b * 649 * , 111 ( 2007 ) . b. w. harris and j. smith , phys . d * 57 * , 2806 ( 1998 ) . s. chekanov _ et al . _ [ zeus collab . ] , arxiv:0704.3562 [ hep - ex ] . s. fang [ zeus collab . ] , `` charm fragmentation function and fragmentation functions of charm mesons at zeus , '' _ these proceedings . _ a. aktas _ et al . _ [ h1 collab . ] , arxiv : hep - ex/0701023 . h. jung , comput . commun . * 143 * ( 2002 ) 100 . a. aktas _ et al . _ [ h1 collab . ] , eur . j. c * 50 * , 251 ( 2007 ) . a. geiser , `` review of beauty production at hera , '' _ these proceedings . _ a. aktas _ et al . _ [ h1 collaboration ] , eur . j. c * 40 * ( 2005 ) 349 , + a. aktas _ et al . _ [ h1 collaboration ] , eur . j. c * 45 * ( 2006 ) 23 . k. krger , `` review on charmonium production , '' _ these proceedings . _ e. l. berger and d. l. jones , phys . d * 23 * , 1521 ( 1981 ) , + r. baier and r. rckl , phys . b * 102 * , 364 ( 1981 ) . f. abe _ et al . _ [ cdf collab . ] , phys . lett . * 79 * , 572 ( 1997 ) , + t. shears [ cdf collab . ] , eur . phys . j. c * 33 * , s475 ( 2004 ) . g. t. bodwin , e. braaten and g. p. lepage , phys . d * 51 * , 1125 ( 1995 ) [ erratum - ibid . d * 55 * , 5853 ( 1997 ) ] . a. abulencia _ et al . _ [ cdf collab . ] , arxiv : hep - ex/0703028 . m. steder [ h1 collab . ] , `` inelastic @xmath44 production in dis at h1 , '' _ these proceedings . _ m. krmer , nucl . phys . b * 459 * , 3 ( 1996 ) . c. adloff _ et al . _ [ h1 collab . ] , eur . j. c * 25 * , 25 ( 2002 ) , + s. chekanov _ et al . _ [ zeus collab . ] , eur . j. c * 27 * , 173 ( 2003 ) . m. zur nedden [ hera - b collab . ] , `` recent charmonium results from hera - b , '' _ these proceedings . _ a. adare _ et al . _ [ phenix collab . ] , phys . lett . * 98 * , 172301 ( 2007 ) . b. i. abelev _ et al . _ [ star collab . ] , arxiv : nucl - ex/0607012 . e. t. atomssa [ phenix collab . ] , `` @xmath44 suppression measurements by the phenix experiment at rhic , '' _ these proceedings . _ s. k. choi _ et al . _ [ belle collab . ] , phys . * 89 * , 102001 ( 2002 ) [ erratum - ibid . * 89 * , 129901 ( 2002 ) ] , + d. m. asner _ [ cleo collab . ] , phys . rev . lett . * 92 * , 142001 ( 2004 ) , + k. abe _ et al . _ [ belle collab . ] , arxiv : hep - ex/0507019 . j. l. rosner _ [ cleo collab . ] , phys . lett . * 95 * , 102003 ( 2005 ) . w. m. yao _ et al . _ [ particle data group ] , j. phys . g * 33 * , 1 ( 2006 ) . c. cawlfield _ et al . _ [ cleo collab . ] , phys . * 98 * , 092002 ( 2007 ) . v. poireau [ belle and babar collab . ] , `` new resonances and meson spectroscopy at belle and babar , '' _ these proceedings . _ g. gokhroo _ et al . _ [ belle collab . ] , phys . lett . * 97 * , 162002 ( 2006 ) . b. aubert _ et al . _ [ babar collab . ] , arxiv : hep - ex/0607083 . t. e. coan _ et al . _ [ cleo collab . ] , phys . rev . * 96 * , 162003 ( 2006 ) . q. he _ et al . _ [ cleo collab . ] , phys . rev . d * 74 * , 091104 ( 2006 ) . b. aubert _ et al . _ [ babar collab . ] , phys . lett . * 95 * , 142001 ( 2005 ) . k. abe _ et al . _ [ belle collab . ] , arxiv : hep - ex/0612006 . g. pakhlova [ belle and babar collab . ] , `` charm physics at b factories , '' _ these proceedings . _ k. abe _ et al . _ [ belle collab . ] , phys . lett . * 98 * , 092001 ( 2007 ) . h. mendez [ cleo collab . ] , arxiv : hep - ex/0702008 . b. aubert _ et al . _ [ babar collab . ] , arxiv : hep - ex/0610057 . k. abe _ et al . _ [ belle collab . ] , arxiv : hep - ex/0507019 , + k. abe _ et al . _ [ belle collab . ] , phys . lett . * 94 * , 182002 ( 2005 ) , + s. uehara _ et al . _ [ belle collab . ] , phys . lett . * 96 * , 082003 ( 2006 ) . b. aubert _ et al . _ [ babar collab . ] , phys . rev . * 97 * , 222001 ( 2006 ) . b. aubert _ et al . _ [ babar collab . ] , phys . lett . * 98 * , 012001 ( 2007 ) , + k. abe _ et al . _ [ belle collab . ] , arxiv : hep - ex/0608043 . r. chistov _ et al . _ [ belle collab . ] , phys . lett . * 97 * , 162001 ( 2006 ) , + b. aubert _ et al . _ [ babar collab . ] , arxiv : hep - ex/0607042 . b. aubert _ et al . _ [ babar collab . ] , phys . * 97 * , 232001 ( 2006 ) . m. heck [ cdf and d0 collab . ] , `` recent results on b spectroscopy at the tevatron , '' _ these proceedings . _ k. abe _ et al . _ [ belle collab . ] , phys . lett . * 96 * , 221601 ( 2006 ) . e. barberio _ et al . _ [ heavy flavor averaging group ( hfag ) ] , arxiv:0704.3575 [ hep - ex ] . e. barberio _ et al . _ [ heavy flavor averaging group ( hfag ) ] , arxiv : hep - ex/0603003 . m. misiak _ et al . _ , phys . lett . * 98 * , 022002 ( 2007 ) , + t. becher and m. neubert , phys . rev . * 98 * , 022003 ( 2007 ) . k. abe _ et al . _ [ belle collab . ] , phys . lett . * 96 * , 221601 ( 2006 ) , + b. aubert _ et al . _ [ babar collab . ] , phys . lett . * 98 * , 151802 ( 2007 ) . _ et al . _ [ belle collab . ] , phys . * 97 * , 251802 ( 2006 ) , + b. aubert _ et al . _ [ babar collab . ] , arxiv : hep - ex/0608019 . m. corcoran [ cdf and d0 collab . ] , `` b hadron rare decays and lifetimes , '' _ these proceedings . _ r. kehoe [ cdf and d0 collab . ] , `` @xmath171 mixing and lifetime difference measurements , '' _ these proceedings . _ v. m. abazov _ et al . _ [ d0 collab . ] , phys . lett . * 97 * , 021802 ( 2006 ) . a. abulencia _ et al . _ [ cdf collab . ] , phys . lett . * 97 * , 062003 ( 2006 ) . v. m. abazov _ et al . _ [ d0 collab . ] , phys . * 98 * , 121801 ( 2007 ) . a. lenz and u. nierste , arxiv : hep - ph/0612167 . v. m. abazov _ et al . _ [ d0 collab . ] , arxiv : hep - ex/0702049 . l. m. zhang _ [ belle collab . ] , phys . lett . * 96 * , 151801 ( 2006 ) , + b. aubert _ et al . _ [ babar collab . ] , phys . * 97 * , 221803 ( 2006 ) . b. aubert _ et al . _ [ babar collab . ] , phys . rev . lett . * 98 * , 211802 ( 2007 ) , + m. staric _ et al . _ [ belle collab . ] , phys . * 98 * , 211803 ( 2007 ) . k. abe _ et al . _ [ belle collab . ] , arxiv:0704.1000 [ hep - ex ] . a. krasznahorkay jr . [ atlas and cms collab . ] , `` outlook for b and c physics at the lhc in atlas and cms , '' _ these proceedings . _ s. j. brodsky , p. hoyer , c. peterson and n. sakai , phys . b * 93 * , 451 ( 1980 ) . f. s. navarra , m. nielsen , c. a. a. nunes and m. teixeira , phys . d * 54 * , 842 ( 1996 ) . w. melnitchouk and a. w. thomas , phys . b * 414 * , 134 ( 1997 ) . j. pumplin , h. l. lai and w. k. tung , phys . d * 75 * , 054029 ( 2007 ) _ and these proceedings . _ m. krmer , j. zunft , j. steegborn and p. m. zerwas , phys . b * 348 * , 657 ( 1995 ) . m. klasen , b. a. kniehl , l. n. mihaila and m. steinhauser , nucl . b * 713 * , 487 ( 2005 ) . m. klasen , b. a. kniehl , l. n. mihaila and m. steinhauser , phys . d * 71 * , 014016 ( 2005 ) . j. campbell , f. maltoni and f. tramontano , arxiv : hep - ph/0703113 . b. a. kniehl , d. v. vasin and v. a. saleev , phys . d * 73 * , 074022 ( 2006 ) . b. a. kniehl , v. a. saleev and d. v. vasin , phys . d * 74 * , 014024 ( 2006 ) _ and these proceedings . _ j. j. dudek and r. g. edwards , phys . lett . * 97 * , 172001 ( 2006 ) . lansberg and t. n. pham , phys . d * 74 * , 034001 ( 2006 ) j. p. lansberg and t. n. pham , phys . d * 75 * , 017501 ( 2007 ) . j. j. dudek , r. g. edwards and d. g. richards , phys . d * 73 * , 074507 ( 2006 ) . y. j. gao , y. j. zhang and k. t. chao , chin . * 23 * , 2376 ( 2006 ) . y. j. gao , y. j. zhang and k. t. chao , arxiv : hep - ph/0701009 . j. m. richard , nucl . * 164 * , 131 ( 2007 ) _ and these proceedings .
during the last year many important results have been achieved in heavy flavour physics : new measurements of charm and beauty production have been performed at hera and the tevatron . a wealth of new spectroscopy data with several new , unexpected states in the charmonium and the @xmath0 systems has been collected and @xmath1 transitions have been established . the oscillation frequency in the @xmath2 is now measured , and mixing in the @xmath3 system has been observed . theoretical progress in the areas of open heavy flavour production , quarkonium production and decays , and multiquark spectroscopy has been presented at this workshop . lpsc 07 - 73 * summary of the heavy flavor working group * michael klasen@xmath4 , benno list@xmath5 , stephanie hansmann - menzemer@xmath6 , and rainer mankel@xmath7 + 1- universit grenoble i - + laboratoire de physique subatomique et de cosmologie + 53 avenue des martyrs - 38026 grenoble - france + 2- universitt hamburg - institut fr experimentalphysik + luruper chaussee 149 - 22761 hamburg - germany + 3- universitt heidelberg - physikalisches institut + philosophenweg 12 - 69120 heidelberg - germany + 4- deutsches elektronen - synchrotron desy - f1 + notkestr . 85 - 22603 hamburg - hamburg to appear in the proceedings of the xv international workshop on deep - inelastic scattering and related subjectes , dis 2007 , april 1620 , 2007 , munich , germany .
the dust and gas that comprise the interstellar medium ( ism ) play an important role in galaxy evolution , despite the fact that they constitute a relatively small fraction of each galaxy s total mass ( e.g. , @xmath710% in the milky way ) . the presence of interstellar dust is physically manifested both through absorption and scattering processes which redistribute @xmath630% of the starlight emitted at uv - optical wavelengths over a range of infrared ( ir ) wavelengths @xcite , and through the depletion of refractory elements from ism gas . the physical and chemical states of the ism are significantly affected by the dust , which regulates the thermal energy balance through the ejection of photoelectrons in photodissociation regions , and via ir emission from molecular and shock - heated plasmas . dust grain surfaces also promote the formation of molecules , such as h@xmath8 , which act as a coolant facilitating molecular cloud collapse . as the ism gas is formed into new stars and the galaxy s stellar population evolves , the ratio of the silicate dust to the carbonaceous dust reflects the relative abundances of their different stellar sources . silicate dust is typically produced in oxygen - rich environments , such as near type ii supernovae or agb stars with low c / o ratios , while carbonaceous dust is associated with carbon - rich environments , such as near evolved low - mass stars , although the details of the formation are still debated both in local and in higher redshift galaxies [ e.g. , @xcite ] . since dust is relatively quickly destroyed through shock processing in the ism , with a typical lifetime of @xmath9500 myr for many dust species @xcite , interstellar dust is a sensitive probe of the state of the ism material and stellar populations at a given epoch , assuming that selective destruction and reformation rates in the ism are not highly variable . quasar absorption systems ( qass ) provide a unique tool to simultaneously probe the dust and gas content in the ism of distant galaxies , independent of galaxies intrinsic luminosities or star - formation rates . the study of qass exploits sightlines to distant , luminous objects , such as quasars , that emit radiation over a wide - range of frequencies . as the quasar light passes through foreground galaxies , interstellar gas and dust produce absorption features in the quasar spectrum . the ism of any galaxy with a luminous background source can be studied with this technique , but it is particularly effective when applied to the most gas - rich systems , such as damped lyman-@xmath10 absorbers ( dlas ; neutral hydrogen column densities n@xmath11 @xmath12 ) and sub - dlas ( 10@xmath13n@xmath14 @xmath12 ) . these gas - rich qass are the primary neutral gas reservoir for star formation in the universe [ e.g. , @xcite ] . as opposed to flux - limited studies of gas and dust in emission , that are biased toward the most robustly star - forming objects , qass provide insight into more quiescent or moderately - star - forming galaxies , potentially more similar to local group galaxies . evidence for the presence of interstellar dust in gas - rich qass has been found in both moderate - to - large surveys , as well as in investigations of individual systems . studies of qass show depletions of refractory elements and reddening of background quasars @xcite , as well as a correlation between dust - reddening and gas metal absorption line strengths in absorbers at 1@xmath15z@xmath152 @xcite . the bulk of analyses exploring dust spectral properties in qass , however , concentrate on signatures associated with the carbonaceous dust component , such as the 2175 bump , rather than on the silicate dust , which exhibits prominent ir features near 10 and 18 @xmath1 [ see , e.g. , @xcite for review of dust spectral signatures and their origins ] . it is important to also consider this silicate dust , both because it is primarily produced by a different stellar sub - population , and because within the milky way @xmath766% of the core mass of interstellar dust grains is silicate in composition @xcite . in an exploratory study , our group made the first measurement of silicate dust in a qas . a @xmath610-@xmath2 detection of the 9.7 @xmath16 m silicate absorption feature was made in the z@xmath17 dla toward ao 0235 + 164 using low - resolution _ spitzer space telescope infrared spectrograph ( irs ) _ data . this was followed by identification of the 9.7 @xmath1 silicate absorption feature in 5 additional strong qass at redshifts [email protected] @xcite . analyses of the silicate dust in these 6 qass suggest both similarities and differences relative to that in the milky way ism . in the milky way diffuse interstellar clouds , the peak optical depth of the silicate absorption is seen to scale with the extinction ( @xmath19 ) , and hence with the reddening [ e(b - v)=@xmath20 where @xmath21 , the selective extinction , is @xmath73.1 ] . in the studied qass , the silicate optical depth similarly scales with the reddening . however , the absorption features are 2 - 4 times deeper for a given reddening than those in the milky way @xcite . this disparity may stem from differences in dust composition , metallicity , and/or grain size distributions . a larger selective extinction may imply , for instance , larger dust grain sizes . furthermore , in these 6 qass comparisons of the peak wavelengths of the 10 @xmath1 silicate absorption feature , and of the substructure within the feature , reveal variations suggestive of mineralogical and grain structural differences relative to one - another , and relative to the milky way ism @xcite . these differences are most pronounced in the z=0.886 absorber toward pks 1830 - 211 @xcite , where the 10 @xmath1 absorption feature substructure is consistent with a crystalline olivine silicate dust grain origin , in marked contrast with the amorphous ( @xmath155% crystalline ) silicate dust grains in the milky way ism @xcite . exploring whether these trends and variations are found in other qass is essential for quantifying the evolution in dust grain properties . in this paper , we make progress toward this goal by presenting the first analysis of the 10 and 18 @xmath1 silicate absorption features in the z@[email protected] absorber @xcite toward the gravitationally lensed , [email protected] blazar txs 0218 + 357 at ra ( j2000)= 0221055 and dec ( j2000 ) = + 355614 @xcite . throughout our analysis we simply refer to this z@xmath0=0.6847 system , which is by far the strongest known metal absorption system along the quasar sightline , as _ the _ txs 0218 + 357 qas . the txs 0218 + 357 qas was selected for this study for several reasons . first , this qas is relatively rich in both gas and dust . the abundance of gas is evident from the large measured mg ii absorption rest - frame equivalent width [email protected] , the high molecular hydrogen column density ( @xmath23 @xmath12 ) , and the detections of 21-cm absorption implying @xmath24 @xmath12 , where @xmath25 is the gas spin temperature and @xmath26 is the h i covering factor @xcite . there are also signatures of dust in the qas , including an exceedingly high measured reddening of the background quasar . statistical studies of mean dla extinction curves using sdss data find that most qass at 1@xmath15z@xmath155 have a typical rest - frame reddening of e(b - v)@xmath70.002 @xcite , while the txs 0218 + 357 qas exhibits a total measured e(b - v)[email protected] , assuming r@xmath27 @xcite . second , this qas is known to be rich in molecules , similar to the pks 1830 - 211 z=0.886 absorber . since in the pks 1830 - 211 qas the interstellar silicate dust may be significantly crystalline @xcite , unlike that in the milky way , it is important to ascertain whether other molecule - rich qass are likewise potentially rich in crystalline silicates . molecules such as co , ch , and h@xmath8 trace cold , dense , star - forming material in the galaxy ism . the txs 0218 + 357 qas exhibits transitions for numerous molecules including co , hco@xmath28 , hcn , hnc , cn , cs , c@xmath29o , @xmath30co , h@xmath8o , nh@xmath31 , oh , h@xmath8co and its anti inversion , and the primordial lih molecule which traces h@xmath8 in low metallicity regions @xcite . moreover , this absorber has among the highest molecular column densities at this redshift range @xcite . studies of the qas reveal that one of the gravitationally - lensed quasar sightlines passes through an optically thick and clumpy molecular cloud , while the other sightline passes through a more optically thin region @xcite . the derived column densities and abundances for molecules such as formaldehyde along the more obscured sightline are similar to those found in milky way molecular clouds @xcite , while the nh@xmath31 absorption measurements are more consistent with a diffuse cloud origin @xcite . third , the z@xmath0=0.68466 qas redshift is optimal for studying both the 10 and 18 @xmath1 silicate absorption features with the _ irs _ instrument used in our previous studies . in the qas rest - frame , the irs spectra span the region 3.1 - 22.6 @xmath1 , with adequate quasar continuum coverage between the absorption features additionally , for this quasar the _ irs _ sensitivity was sufficient to obtain data for this analysis in only 30 minutes of exposure time . fourth , the z@xmath32=0.944 @xcite background quasar is gravitationally lensed , which boosts its signal , and hence the signal - to - noise ratio ( s / n ) of the qas gas and silicate dust absorption features . furthermore , since this is a well - studied strong lens system , information about the qas host galaxy topography , which is often lacking for other qass , is available . studies of the morphology of the lensing galaxy hosting the qas indicate it is a gas - rich spiral galaxy with narrow absorption ( ca ii h&k , mg ii , fe ii , mg i ) and emission ( [ oii ] , h@xmath33 , [ oiii ] ) lines @xcite , suggestive of some ongoing star - formation , comparable to that in spiral galaxies like the milky way based on the measured [ oii ] equivalent width @xcite . radio studies [ e.g. , @xcite ] reveal that the quasar is being lensed into two point sources and a small einstein ring with an angular diameter of 335 mas ( @xmath72.4 kpc in the lensing galaxy for h@xmath34=70 km / s / mpc ) . the optical and ir magnitudes along the two compact - object lines of sight imply a reddened and optically faint ( v@xmath621mag ) source @xcite . lastly , owing to the well - studied strong gravitational lens , as well as to the monitored ( variable ) blazar , abundant ancillary data are available for this system . these data include spectral energy distributions ( seds ) extending from the @xmath35 ray ( e.g. fermi , rosat , swift ) through the optical and nir ( e.g. hst ) , to the sub - mm ( e.g. nrao , pdbi ) and radio ( e.g. , vla , vlba ) . however , while these data provide global constraints , owing to the blazar variability , predicting the flux at a given epoch is challenging . in this paper , we begin with a description of the _ irs _ observations , followed by a summary of the data reduction ( [ irsspect ] ) , spectral extractions ( [ extract ] ) , and quasar continuum normalizations ( [ norm ] ) . we then discuss the profile template fitting to the silicate absorption features produced by the txs 0218 + 357 qas ( [ results ] ) . we conclude with a discussion of our findings in comparison with the silicate dust in other qass ( [ disc ] ) , followed by a summary of our main results ( [ sum ] ) . in the appendices , we show the effects of different quasar continuum normalizations on our results ( appendix [ normap ] ) , and discuss details of the template profiles applied to the data ( appendix [ templateap ] ) . our data consist of spectra of the txs 0218 + 357 quasar obtained using the _ irs _ instrument aboard the _ spitzer space telescope _ , as summarized in table [ obsit ] . these data were observed as part of program 50783 ( pi : v.p . kulkarni ) in the irs staring mode . the blue peak - up filter with high peak - up accuracy was used to identify the quasar . five exposures were taken in each of the two nod positions at every available low - resolution spectral order , for a total of 10 exposures per order . lccc sl2 & 5.2 - 7.7 & 3.1 - 4.6 & 10x60s + sl3 & 7.3 - 8.7 & 4.3 - 5.2 & 10x60s + sl1 & 7.4 - 14.5 & 4.4 - 8.6 & 10x60s + ll2 & 14.0 - 21.3 & 8.3 - 12.6 & 10x30s + ll3 & 19.4 - 21.7 & 11.5 - 12.9 & 10x30s + ll1 & 19.5 - 38.0 & 11.6 - 22.6 & 10x30s + [ obsit ] although the quasar txs 0218 + 357 is gravitationally lensed into a 335 mas diameter einstein ring and two bright , compact components also separated by 335 mas @xcite , our data do not resolve these individual sightlines through the foreground absorber . according to the irs manual , the fwhm spatial profile widths obtained for calibration point sources ( stars ) exceed 25 for every spectral order . _ irs _ slits range in width from 36 ( sl2)-107 ( ll1 ) and in length from 57 - 168@xcite . each _ irs _ spectrum , thus , is the superposition of the multiple sightlines through the qas , and so we can not investigate spatial variations in the galaxy ism . the 2-dimensional ( 2d ) spectra obtained over the 40 total exposures were automatically reduced using the _ spitzer irs _ basic calibrated data ( bcd ) pipeline version s18.18.0 . this pipeline applied bias and dark subtractions , as well as flat - fielding the data . it also corrected for systemic effects such as droop and dark - current drift @xcite . prior to subtracting the background and extracting the quasar spectrum , we implemented two phases of bad - pixel cleaning ( i.e. replacement ) . the background bad pixels were first cleaned using _ irsclean _ ( version 2.1 ) , followed by the manual replacement of remaining anomalous background pixels and hot / cold pixels along the quasar spectrum using the _ image reduction and analysis facility ( iraf ) epix _ task . typically fewer than 6 pixels were replaced along the quasar spectrum in a single exposure . our tests find that our cleaning method results in a more conservative estimate of the significance of the structural parameters than would be obtained using the uncleaned , raw bcd data . the cleaned exposures in each nod position were combined using the _ spitzer _ _ idl _ script _ coad_. each pixel value in the combined frame was computed as the trimmed mean of the pixels in the input images . pixels with no value and those fatally flagged in the associated mask files were excluded . the corresponding pixel uncertainty was computed as the square root of the sum of the squares in the input uncertainty file , normalized by the number of pixels used to calculate the trimmed mean . we subtracted the background from each of the coadded frames using the nod exposures . a background - subtracted 2d spectrum was produced by subtracting the nod frames from one another . the associated uncertainty files were propagated in quadrature . the wavelength - calibrated , 1d quasar spectrum was extracted from the background - subtracted , 2-d spectrum for each nod using the _ spice gui _ ( version 2.5 ) . each extraction was run using the generic point - source extraction mode with the optimal extraction setting . the optimal extraction weights each pixel in the aperture using the s / n of the pixel , as computed using the associated pipeline uncertainty file . this method improves the s / n for intrinsically low - s / n spectra . the spectra from the two nods were combined by averaging the flux values at every wavelength . the associated uncertainties are a combination in quadrature of the formal uncertainty associated with the mean ( determined by propagating the pipeline uncertainties ) and the scatter of the flux measurements about the mean at a given wavelength . in general , data points in the final spectrum with large error bars stem from discrepancies between the two nod spectra . these differences may originate from a slightly hot pixel along the quasar spectrum or from an emission feature which lies on a cold pixel in one nod . rather than eliminating these data points or simplistically assuming them to originate in hot pixels , we conservatively retain them . these noisy points have a minimal impact on our absorption feature fitting , because their large uncertainties effectively de - weight the data points . the nod - averaged spectra ( sl2 , sl1 , ll2 , ll1 ) were stitched together to produce a single contiguous spectrum . a few ( @xmath363 ) anomalous data points at each spectral segment edge were removed to produce a smoother joining . all data at @xmath3735.8 @xmath1 were rejected , as they showed significant degradation . the sl2 and sl1 data were scaled by multiplicative factors to match the identically cleaned and extracted overlap ( sl3 ) data . a similar process was repeated for scaling the ll1 and ll2 orders to match the the ll overlap ( ll3 ) data , and scaling the ll orders to match the sl orders . the final multiplicative rescaling factors for the sl2 , sl1 , ll2 , and ll1 orders were 0.82 , 0.92 , 1.14 , and 1.05 , respectively . since the resultant quasar spectrum is continuum normalized prior to our analysis of the silicate absorption features , as described below , the absolute flux normalization is not crucial . the overlapping data points at the order joinings were averaged to produce a single flux at each wavelength . the errors are a combination of the propagated measurement uncertainties and the standard deviation about the mean for each data point . a visual inspection indicated that the uncertainties in the sl1/sl2 and ll1/ll2 overlap regions were typically larger than those in the associated overlap spectra ( sl3 , ll3 ) . therefore , in regions where an overlap spectrum flux value was available , it was used in place of the first and second order data . tests indicate that these replacements slightly reduce the mean uncertainties in the final spectrum , while not impacting the shape of the absorption features or quasar continuum . the final combined spectrum is shown in figure [ spect ] ( left ) , as a function of the observed wavelength . we do not bin the data in this analysis , as was done in the analyses of several previous qass @xcite , because while binning reduces the data point uncertainties , it may also obscure substructure in the absorption features . the combined , unbinned quasar spectrum was finally shifted to the z=0.68466 redshift qas rest - frame . this qas rest - frame spectrum was used in all subsequent steps of the analysis , including fitting the quasar continuum and performing the mineralogical fits to the silicate absorption features . all wavelengths referred to in the following sections are in the qas rest - frame , unless explicitly stated otherwise . we normalized the combined spectrum by the quasar continuum , prior to measuring the silicate absorption features . the physical shape of the quasar continuum is unknown . therefore , we fitted a series of low - order functions to our data , including linear , power - law , 3rd order chebyshev polynomial , and cubic spline fits . we experimented with several different fitting regions and data point exclusions for each fit , as detailed in appendix [ normap ] . we selected as the best - fitting a 3rd order chebyshev polynomial fitted by excluding the 8.5 - 12.5 @xmath1 and 16.5 - 20.4 @xmath1 regions containing the silicate absorption features , and manually rejecting several data points with anomalously large uncertainties and deviant flux values in the ll1/ll2 joining region . this fit visually matches the data slightly better than the other considered quasar continuum fits , although _ all _ normalizations result in similar 10 @xmath1 silicate feature fits , as detailed in appendix [ normap ] . the largest discrepancies occur for the 18 @xmath1 absorption feature , which is near the ` red ' end of our covered spectral range . depending on the shape of the adopted fit , the 18 @xmath1 feature may be relatively broad and deep , or may be narrow and shallow . since there are no mid - ir ( _ spitzer irac , mips _ ; _ herschel pacs _ ) archival data to constrain the slope of the continuum beyond our spectrum , and far - ir / radio data obtained in different epochs can not be compared reliably ( due to the temporal variability of the blazar ) , we can not determine the physical quasar continuum shape near 18 @xmath1 . therefore , we do not emphasize our analysis of the 18 @xmath1 absorption feature . the final quasar - continuum - normalized spectrum we obtained is depicted in figure [ spect ] ( right ) . an examination of the continuum normalized spectrum ( figure [ spect ] , right ) reveals a detection of both relatively strong 10 @xmath1 silicate absorption , as well as a broader and shallower 18 @xmath1 silicate feature . we measured the rest - frame equivalent width ( ew ) of the 10 @xmath1 feature to be [email protected] @xmath1 , which corresponds to a 10.7@xmath2 detection significance . we conservatively estimate the equivalent width uncertainty to be the combination ( in quadrature ) of the photon noise and the continuum noise , with the continuum noise dominating . to estimate the continuum uncertainty , we used the prescription of @xcite , with a nudge factor " of 0.3 . this equivalent width is nearly quadruple that of the 10 @xmath1 absorption feature in the z=0.524 qas toward ao 0235 + 164 , and is almost double that of the largest 10 @xmath1 equivalent width qass previously studied @xcite . we similarly measured the ew of the 18 @xmath1 silicate feature to be [email protected] @xmath1 ; a 4.1@xmath2 significance detection . we caution , however , that the continuum normalization near 18 @xmath1 remains uncertain because of the absence of longer - wavelength data . this feature may be over- or under - estimated . if we assume that the absorption profile is produced by amorphous olivine , as we discuss below , then this equivalent width is a lower limit , because the predicted 18 @xmath1 feature extends past the ` red ' edge of our data . in order to determine the peak optical depth and to place some constraints on the mineralogical composition and grain properties of the dust producing the silicate absorption features , we fitted template profiles to the normalized quasar spectrum . we follow the methodology outlined in @xcite . the transmitted flux at any frequency @xmath38 is given by f@xmath39/f@xmath40 $ ] , where @xmath41 is the measured optical depth at frequency @xmath38 , assuming a constant covering factor ( @xmath42 ) of 1.0 . the optical depth can be expressed in terms of the peak optical depth ( @xmath43 ) and the template unity - normalized optical depth profile ( @xmath44 ) as @xmath45 . the optimal peak optical depth is determined by minimizing the chi - squared over the specified fitting region . the 1-@xmath2 uncertainty is determined from an interpolation of the high and low @xmath43 values , where @xmath46=1.0 relative to the minimum measured chi - squared . we determine which mineral best fits the observed silicate absorption profile by comparing the reduced chi - squared ( @xmath47 ) values and the chi - squared probabilities , which quantify the probability that a similar or larger @xmath48 would be obtained by chance . we considered a representative sample of silicate template profiles obtained from the literature for laboratory minerals and for astrophysical samples of sic . these laboratory templates include both amorphous olivines ( mg@xmath49fe@xmath50sio@xmath51 ) and pyroxenes ( mg@xmath52fe@xmath53sio@xmath31 ) and crystalline olivines and pyroxenes , with a range of chemical compositions and grain temperatures , shapes , porosities , and sizes . we explore multiple templates in each mineral category because studies [ e.g. , @xcite ] have established that physical variations in these grain properties , as well as the laboratory methodology employed to measure the mineral properties for a given species , can significantly impact the silicate profile shape and peak absorption wavelengths . in our analysis , however , we only describe the results for the best - fitting ( lowest @xmath47 and highest probability ) profile fit in each mineral category . all of the template profiles have been measured for , or utilized in , comparison with galactic astrophysical dust sources , and are discussed in more detail in appendix [ templateap ] , and in @xcite . we emphasize that our goal in exploring these template profile fits is _ not _ to unambiguously determine the mineral and dust grain properties which explain the observed silicate dust absorption . this would require much higher s / n data . rather , we use the templates to place some constraints on the dust grain properties , by exploring whether certain classes of minerals are viable and whether certain classes of minerals can be ruled - out as plausible origins for the observed silicate features . we explore several minerals in each category solely to ensure that an entire mineral category is not ruled out simply because one or two species of minerals in the category fit the data poorly . in addition to considering the single mineral fits , we also briefly consider the possibility that the silicate dust may be a mixture of two different minerals or grain structures . @xcite found that silicate dust in ulirgs was best explained by a mixture of amorphous and crystalline olivines . likewise , there is no reason why the interstellar dust in the txs 0218 + 357 qas must be homogeneous . we implement these two - mineral fits by determining @xmath54 and @xmath55 in the expression f@xmath56/f@xmath57 $ ] again assuming a constant covering factor of @xmath58 , following the methodology described in @xcite we commence with fitting the template profiles to _ only _ the txs 0218 + 357 qas 10 @xmath1 silicate absorption feature . we restrict fitting ( i.e. @xmath48-minimization ) to the @xmath59 region in which the absorption occurs . this relatively narrow fitting region is adopted to prevent undue weighting from other , weaker spectral features which may exist within the quasar continuum . small changes in the breadth of the fitting region have little impact on our results . the fitting results are summarized in table [ fits-10 ] and the fits are illustrated in the left panels of figures [ bestsinglefit]-[singlefit ] . the best fit is produced by a species of amorphous olivine with porous , ellipsoidal dust grains ( amorph.oliv.gpc ) , with a measured peak optical depth for the 10 @xmath1 feature of @[email protected] . as illustrated in the left panel of figure [ bestsinglefit ] , this profile well matches the depth , breadth , and peak location of the silicate feature . the best mineral fits in the other mineral categories are inferior both statistically ( table [ fits-10 ] ) and visually ( figure [ singlefit ] ) . amorphous pyroxene and sic have both been exploited to explain silicate dust absorption in milky way environments such as near agb stars [ e.g. @xcite ] . however , we find , for our data on the txs 0218 + 357 qas , that the best - fitting amorphous pyroxene ( amorph.pryox.gpc ; porous , ellipsoidal grains ) and sic ( green @xmath10-sic ) templates explored produce particularly poor fits to the data ( @xmath48 fit probabilities consistent with 0% ) . the fits neither match the peak wavelength nor the breadth of the observed absorption feature . crystalline olivines and pyroxenes have been utilized to explain silicate absorption features in both galactic [ e.g. , @xcite ] and extragalactic [ e.g. , @xcite ] dust . our best - fitting crystalline olivine ( t=10k low - temperature forsterite ; mg@xmath60sio@xmath51 ) and crystalline pyroxene ( orthobronzite ; mg@xmath61fe@xmath62sio@xmath31 ) fits are mildly suggestive that some crystalline material could be present , but the fits are poorer than for amorphous olivine ( see table [ fits-10 ] ) . although some features expected for these minerals align with substructure in our absorption profile , such as the small central ridge in the 10 @xmath1 absorption feature , all of the substructure is our data is well - within the noise uncertainty limits , and so does not provide compelling evidence for crystallinity . if we instead used the galactic astrophysical profile templates discussed in @xcite to fit the data , our peak optical depth for the 10 @xmath1 feature would be largely unchanged . of the three templates , the diffuse cloud @xmath16cephei red supergiant @xcite template produces the best match to our data , with @xmath63 and a @xmath64 ( p=59% ) . while the optical depth is consistent with that derived using the laboratory amorphous olivine template , the fit quality is poorer . the gcs-3 galactic center diffuse ism source @xcite also provides an adequate fit ( @xmath47=1.09 , p=26% ) with a consistent optical depth ( @xmath65 ) . the only template discussed in @xcite which does not provide a good fit to our data is the trapezium molecular cloud @xcite template which has a @xmath66 fit ( @xmath48 fit probability consistent with 0% ) . based on these astrophysical templates , we would conclude that the txs 0218 + 357 qas is not well matched by galactic molecular clouds , but is broadly consistent with material in other , more diffuse , galactic environments . lllllll amorph . olivine ( best ) & amorph.oliv.gpc ( 1 ) & [email protected] & 0.86 & 81.8 & 85 & 8.02 - 12.10 + amorph . pyroxene & amorph.pyrox.gpc ( 1 ) & [email protected] & 1.77 & 0.0 & 86 & 8.00 - 12.10 + cryst . olivine & forsterite(t=10k)(2 ) & [email protected] & 1.49 & 0.2 & 86 & 8.00 - 12.10 + cryst . pyroxene & bronzite ( 3 ) & [email protected] & 1.08 & 28.7 & 77 & 8.34 - 12.10 + sic & g0-green - asic ( 4 ) & [email protected] & 4.99 & 0.0 & 86 & 8.00 - 12.10 + [ fits-10 ] given the uncertainty associated with the quasar continuum normalization near 18 @xmath1 , we do not explicitly fit template profiles to this feature alone . we instead consider the 18 @xmath1 silicate absorption feature contribution in two ways . first , we examine the predicted shape of the 18 @xmath1 feature based on the derived 10 @xmath1 peak optical depth . second , we consider fits in which the @xmath48-minimization is implemented over @xmath67 , thus simultaneously including the 10 @xmath1 and the 18 @xmath1 features in determining the best mineral fits and peak optical depths . we illustrate the predicted 18 @xmath1 feature , based on our previously described 10 @xmath1 fits , in the central panels of figures [ bestsinglefit]-[singlefit ] . the predicted amorphous olivine 18 @xmath1 absorption feature is a good match to the observed spectrum . the predicted turnover between 10 and 18 @xmath1 follows the data , and both the peak and breadth of the feature are well - aligned with the prediction . this both provides additional support for the presence of amorphous olivine at a measured @xmath68 peak optical depth , and lends credence , a posteriori , to our adopted quasar continuum normalization . once again , sic and amorphous pyroxene provide overall poor predictions of the shapes and locations of observed 10 and 18 @xmath1 features . the crystalline olivine and crystalline pyroxene fits match some of the low - level substructure near 18 @xmath1 , although this substructure is well within the noise limits , but other predicted features such as a rise near 16 - 17 @xmath1 are absent or muted in the data . we next extended the fitting to cover both the 10 and 18 @xmath1 regions simultaneously , the results of which are detailed in table [ fits-10 + 18 ] . these fits are depicted in the rightmost panels of figures [ bestsinglefit]-[singlefit ] . we find that over this expanded fitting region , the best - fitting mineral remains the amorphous olivine species with porous , ellipsoidal grains . the derived peak optical depth of this fit ( @xmath69 ) is consistent with that derived when fitting over only the 10 @xmath1 region . amorphous pyroxene and sic still produce poor fits . the largest differences occur for the crystalline olivines and pyroxenes . different species are identified when the 18 @xmath1 silicate absorption feature is included in the fit : hortonolite ( mg@xmath70fe@xmath71sio@xmath51 ) and natural orthoenstatite ( mg@xmath72fe@xmath73sio@xmath31 ) , respectively . in figure [ singlefit ] we illustrate the species of crystalline olivine and pyroxene which produce superior fits over this extended fitting region in cyan . it is apparent that while these species provide slightly poorer fits over the 10 @xmath1 feature than those originally identified , they are also slightly less structure - rich over the 18 @xmath1 feature , which is more consistent with the observed data . higher s / n and higher resolution data would be required to confirm these features and identify the most viable crystalline species . lllllll amorph . olivine ( best ) & amorph.oliv.gpc ( 1 ) & [email protected] & 0.70 & 99.9 & 179 & 8.02 - 21.17 + amorph . pyroxene & amorph.pyrox.gpc ( 1 ) & [email protected] & 1.34 & 0.1 & 180 & 8.00 - 21.17 + cryst . olivine & hortonolite ( 2 ) & [email protected] & 1.18 & 5.2 & 180 & 8.00 - 21.17 + cryst . pyroxene & nat.enstatite ( 2 ) & [email protected] & 1.00 & 47.4 & 180 & 8.00 - 21.17 + sic & g0-green - asic ( 3 ) & [email protected] & 3.96 & 0.0 & 180 & 8.00 - 21.17 + [ fits-10 + 18 ] lastly , we briefly investigate the possibility that a blend of minerals produces the observed 10 @xmath1 silicate absorption feature . for simplicity , we consider pairings of all of the minerals identified in tables [ fits-10]-[fits-10 + 18 ] , as well as pairings containing several additional amorphous and crystalline olivine species . the results of the two - mineral fitting , performed over only the 10 @xmath1 feature , are summarized in table [ fits - two ] and illustrated in green in figure [ twofit ] . in every panel we include the best - fitting amorphous olivine fit ( in red ) for visual reference . we find that the best overall fit is produced by a combination of crystalline olivine ( san carlos olivine , t=100k ) and crystalline pyroxene ( orothobronzite ) with measured peak optical depths of @[email protected] and @[email protected] , respectively . we obtain a @xmath76 , which is slightly lower than that obtained using the single mineral amorphous olivine ( gpc ) template fit ( @xmath77 ) . this results in a slightly higher ( 97% versus 82% ) probability for the fit . however , as illustrated in figure [ twofit ] , although the crystalline olivine - crystalline pyroxene blend has a relatively high fit probability , it also predicts many small oscillations , or substructure features , within the 10 and 18 @xmath1 absorption features . this substructure can not be ruled out , but we favor the smoother prediction of amorphous olivine , due to the lack of compelling evidence for substructure in the observed profile at the s / n of our data . the addition of crystalline olivine to an amorphous olivine with solid , spherical particles ( amorph.oliv.gs ) also formally improves the quality of the fit over a pure amorphous olivine fit ( 95% versus 82% ) . although , again , without higher s / n data it is impossible to determine whether the associated small substructures being fit have a physical origin or whether they stem from noise in the data . the combination of amorphous pyroxene with amorphous olivine does not lead to an improvement in the fit . lllllllll amorph.+amorph . & amorph.pyrox.gpc ( 1 ) & amorph.oliv.gpc ( 1 ) & 0.02@xmath78 & 0.47@xmath79 & 0.87 & 80.4 & 85 & 8.02 - 12.10 + amorph.+cryst . & amoph.oliv.gs ( 1 ) & sancarl.oliv . ( 2 ) & [email protected] & [email protected] & 0.78&93.2&83&8.12 - 12.10 + cryst.+cryst . ( best ) & bronzite ( 3 ) & sancarl.oliv . ( 2 ) & [email protected] & [email protected] & 0.71 & 97.2 & 77 & 8.34 - 12.10 + [ fits - two ] we find that the best - single mineral fit to the silicate absorption features in the txs 0218 + 357 qas is produced by a laboratory amorphous olivine silicate template with a measured peak optical depth of @xmath68 . this relatively high optical depth is not a product of our adopted best - fitting template shape . all of the mineral fits over both the 10 @xmath1 feature and combined 10 and 18 @xmath1 features with @xmath801 result in a measured @xmath81 . for the two mineral pairings , the combined 10 @xmath1 peak optical depths are also consistently high . this derived peak optical depth for the txs 0218 + 357 qas is significantly larger than those measured for any of the 6 previously studied qass , as illustrated in table [ compare ] . despite the deeper absorption , however , the best - fitting mineral species is consistent with several other qass and with silicate dust grains in the milky way ism . we discuss the implications of this high level of absorption and the dust grain species in the following sections . in comparing the peak optical depth measurements from the different qass we now use both the measured peak optical depth values ( @xmath82 ) , as well as the optical depths corrected ( @xmath83 ) by accounting for the covering factor ( @xmath42 ) . this correction has been implemented using the expression @xmath84 the covering factor quantifies the fraction of the subtended area along the sightline to the continuum source obscured by the dusty material , and in many systems may be less than 1.0 , thus implying a higher physical optical depth than that measured . since direct estimates of the dust covering factor along the sightline are not feasible , we assume that the dust traces either the molecular or h i gas , and has a similar covering factor . for the txs 0218 + 357 qas , both molecular line analyses [ e.g. , @xcite ] and h i studies [ e.g. , @xcite ] find a difference between the measured fluxes for the southwest and northeast lensed quasar images , with a flux ratio of @xmath85 at 106 ghz @xcite . however , the precise value of the gas covering factor is debated in the literature [ e.g. , @xcite ] , and the value at 10 @xmath1 may differ from that at @xmath7100 ghz . it has also been suggested that the h i absorption regions may be more extended than the region producing the molecular lines @xcite . for this analysis we adopt a value for the covering factor of @xmath86 , which is broadly consistent with values found by @xcite for molecular material in this system . this results in a corrected optical depth value of @xmath87 for the txs 0218 + 357 qas . for the z=0.886 qas toward the lensed quasar pks 1830 - 211 , we adopt a lower total covering factor of @xmath88 @xcite , which raises the optical depth to @xmath89 . however , the value of @xmath42 is relatively uncertain [ e.g. , @xcite ] , and a slight difference in the covering factor could have a significant impact . for instance , @xmath90 and @xmath91 would imply @xmath92 and @xmath93 , respectively . for the non - lensed systems , covering factors may also be applicable . for ao 0235 + 164 we use @xmath58 @xcite . for the q0852 + 3435 , q0937 + 5628 , and q1203 + 0634 qass estimates of the covering factor are not available in the literature . although @xcite find a median covering factor of @xmath94 for a sample of [email protected] dlas , this value was determined at low frequencies ( @xmath95 1.1 ghz ) where extended radio structure may be contributing to the emission . since such extended structure is unlikely to be associated with the silicate dust , we instead conservatively adopt a value of @xmath58 for all qass without a published covering factor . for the z=0.437 qas toward 3c 196 , although the 21 cm absorption has a very low covering factor against the agn [ @xmath96 ; @xcite ] , it is not clear that such a low covering factor is appropriate for the dusty qas material , due to the differences between the radio and optical sightlines to the quasar . thus , we again adopt a representative @xmath58 for the 3c 196 qas . given the significant uncertainties and assumptions associated with some of these covering factors , when examining correlations with other dust and gas properties we utilize both the measured and the corrected peak optical depth measurements , rather than relying solely on the covering factor - corrected values . llllllllll ao 0235 + 164 & 0.524 & amorph.oliv.gs ( 1 ) & [email protected] ( 2 ) & 1.0 ( 3,4 ) & [email protected] & [email protected] ( 5 ) & [email protected] ( 6 ) & @xmath970.71 ( 7 ) & 0.71 + 3c 196 & 0.437 & amorph.oliv.2.0 ( 8) & [email protected] ( 2 ) & 1.0 & [email protected] & [email protected] ( 2,9 ) & [email protected] ( 10 ) & 0.05 ( 2,11 ) & + q0852 + 3435 & 1.309 & amorph.oliv.gpc ( 1 ) & [email protected] ( 2 ) & 1.0 & [email protected] & [email protected] ( 2 ) & [email protected] ( 2 ) & 0.14 ( 2,12)&0.14 + q0937 + 5628 & 0.978 & amorph.oliv.gpc ( 1 ) & 0.192@xmath98 ( 2 ) & 1.0 & 0.192@xmath98 & [email protected] ( 2 ) & [email protected] ( 2 ) & & + q1203 + 0634 & 0.862 & amorph.pyrox ( 13 ) & 0.277@xmath99 ( 2 ) & 1.0 & 0.277@xmath99 & [email protected] ( 2 ) & [email protected] ( 2 ) & & + pks 1830 - 211 & 0.886 & hortonolite ( 14 ) & [email protected] ( 15 ) & 0.35 ( 16,17 ) & 1.120@xmath100 & [email protected](18 ) & & 0.055 ( 19 ) & 0.17 + txs 0218 + 357 & 0.685 & amorph.oliv.gpc ( 1 ) & [email protected] & 0.8 ( 20,21 ) & 0.667@xmath101 & [email protected] ( 18)&3.0 ( 22 ) & 0.048 ( 23 ) & 0.06 + [ compare ] silicate dust grains in the ism produce 2 prominent spectral features , which we see in the txs 0218 + 357 qas : a 10 @xmath1 and an 18 @xmath1 feature . the former is produced from an si - o stretching mode in the dust molecules , while the latter originates in an o - si - o bending mode . the amount of absorption produced by these two features , i.e. the ratio of their peak optical depths , varies with changes in dust grain properties such as the distribution of grain shapes and sizes . analyses of the ism in local group galaxies [ e.g. , @xcite ] , as well as those in higher redshift systems such as agn tori [ e.g. , @xcite ] , have utilized the observed ratio of these features to constrain dust grain properties . in the txs 0218 + 357 qas , we find the peak optical depth ratio calculated from the observed flux ratios to be @xmath102 . the best - fitting ( lowest @xmath47 ) amorphous olivine template , which had porous , ellipsoidal particles @xcite , has a @xmath103 ratio of 1.41 , which more closely matches the data than , for example , a similar amorphous olivine template with solid , spherical particles that predicts a @xmath103 ratio of 1.62 . ( however , it is not possible to definitively rule out solid , spherical particles given the large uncertainty in the @xmath103 ratio inferred from the data . ) while in principle , we could extend our analysis to compare our observed silicate dust opacity curves with models constructed for different grain temperatures and geometries [ e.g. @xcite ] , we have not taken this step because of the low s / n of our data , and the uncertainties associated with our quasar continuum normalization . as discussed in appendix [ normap ] , since our data do not extend past 21.5 @xmath1 in the qas rest - frame , the shape of the quasar continuum underlying the 18 @xmath1 silicate feature is poorly constrained . differing assumptions about the physical shape of the quasar continuum can significantly affect the depth of the 18 @xmath1 feature . we , therefore , feel that a rigorous comparison with the dust grain models is not warranted for this dataset . given the high peak optical depth measurement for the silicate dust in the txs 0218 + 357 qas relative to the 6 previously studied qass , we have revisited the suggested trend between silicate absorption peak optical depth and reddening presented in @xcite . in that work , it was noted that the trend seemed similar to that for milky way diffuse ism clouds , i.e. @xmath104 for r@xmath105=3.1 , but that the slope differed such that the ratio of @xmath106 was 2 - 4 times that of diffuse galactic clouds . this was used to argue for either a greater abundance of silicate dust in these qass than is found in the milky way ism dust , possibly stemming from differences in the underlying stellar populations , or for differences in the dust grain geometrical and size properties which would affect r@xmath105 . the addition of the txs 0218 + 357 qas and the z=0.886 pks 1830 - 211 qas to the 5 previously studied qass reveals a clear trend of increasing reddening with increasing silicate measured peak optical depth , as illustrated in figure [ ebvplot ] ( left ) . the total reddening measurements are from the literature , as detailed in table [ compare ] . considering all 7 qas we find a pearson linear correlation coefficient of 0.89 ( 0.8% probability of no correlation ) , a spearman ( nonparametric ) rank correlation coefficient of 0.89 ( significance 0.007 , i.e. , 0.7% probability of no correlation ) , and a ( nonparametric ) kendall s tau value of 0.71 ( significance 0.02 ) . the small probabilities / significances indicate a relatively strong correlation . we fitted a linear least - squares relationship to the 7 qass , weighting by both the uncertainties in the reddening from the literature and by those for our measured peak silicate optical depths . we find a slope of [email protected] , which is substantially steeper than the 0.17 slope found for the extrapolation for milky way diffuse clouds from @xcite . our fit is illustrated in figure [ ebvplot ] ( left ) with a solid black line , while the shallower milky way extrapolation is shown as a blue dot - dashed line . all of the studied qass lie significantly above the milky way extrapolation . we note that the steepness of the slope we obtain is being driven in part by the txs 0218 + 357 qas . however , even excluding that qas gives a slope of [email protected] ( dashed red line ) , which is much steeper than the milky way extrapolation . if we require our relationship to extrapolate as a smooth linear function to dust - free systems , i.e. to predict an optical depth of 0.0 for no reddening , we would get a still relatively steep slope of [email protected] ( shown as dotted orange line in figure [ ebvplot ] ) . while this slope is closer to that predicted from the milky way , it remains nearly three times higher than the slope for the extrapolated milky way relation , and the txs 0218 + 357 qas would lie significantly above the relationship . since the reddening error bars are sensitive to the adopted intrinsic quasar spectrum in the extinction - curve - fitting reddening measurements [ e.g. , @xcite ] we have also examined the slope obtained if we omit the reddening uncertainties when fitting over all 7 qass . we find a slope of [email protected] , which does not alter our conclusions . we have also explored the relationship between the reddening and the silicate peak optical depth using the covering factor - corrected peak optical depths , as depicted in figure [ ebvplot ] ( right ) . since the covering factors for several of the qass are highly uncertain , we place less confidence in the derived slope of the relationship , although there is still a general trend of increasing silicate peak optical depth with increasing reddening . furthermore , since the optical depth values are increased by applying the covering factor correction , the offset relative to the milky way diffuse cloud extrapolation increases for @xmath107 . the pearson correlation coefficient is 0.76 ( 5% probability of no correlation ) , with a spearman ( nonparametric ) rank correlation coefficient of 0.89 ( significance 0.007 ) , and a ( nonparametric ) kendall s tau value of 0.71 ( significance 0.02 ) . the magenta dash - dot - dot - dotted line depicts the linear least - squares fit ( slope of [email protected] ) obtained when excluding the txs 0218 + 357 and pks 1830 - 211 qass . when corrected for the covering factors , the molecule - rich pks 1830 - 211 and txs 0218 + 357 qass lie above the trend suggested by the other qass , i.e. , their silicate peak optical depths are higher than expected given their reddening values . this is particularly interesting in the context of studies in the milky way [ e.g. , @xcite ] , which find that while there is a tight correlation between the 10 @xmath1 silicate peak optical depth and the near - infrared color excess @xmath108 along diffuse ism sightlines , when molecular clouds are examined the correlation is absent , with molecular sightlines exhibiting _ lower _ silicate optical depths for a given @xmath108 . the reason for this difference between diffuse and molecular cloud sightlines is still not completely clear . it has been argued that the origin is an increase in @xmath108 in molecular clouds , possibly linked to grain growth or variations in dust species abundances @xcite . our molecule - rich systems , however , lie above the trend established by the other qass . this is more similar to the galactic center sightline , which exhibits a higher silicate optical depth for its observed @xmath108 relative to other galactic diffuse ism sightlines , possibly because of fewer c - rich stars near the galactic center . @xcite . overall , the disparity between the milky way extrapolation and the studied qass seems most significant in systems with large silicate optical depths ; these systems drive the steeper derived slope relative to the milky way . this may indicate that in distant galaxies at relatively high silicate dust columns , the selective extinction differs from that in the milky way because of differing dust grain properties . one possible explanation is that the interstellar dust grains may be larger than those in the milky way ism . large dust grains would yield a low uv extinction , which would translate into a high r@xmath105 , consistent with the measured [email protected] for the txs 0218 + 357 qas @xcite . this decrease in uv extinction , which is not accompanied by a similar decrease in ir absorption , could produce the offset relative to the milky way extrapolation . furthermore , we note that reddening measurements in the milky way sample an integrated combination of different dust grain populations along the galactic plane sightline , while systems such as the txs 0218 + 357 qas are viewed nearly face - on @xcite . this difference in viewing angle may mean that a single , less common , type of dust grain could disproportionately impact the derived total reddening for face - on galaxies . however , the good agreement between our silicate absorption profile and that for amorphous olivine , which has also been used to describe milky way ism silicate dust grains , would imply that the silicate dust grains are somewhat similar . this might suggest that the disparity stems instead from a difference in the underlying stellar populations producing the dust . a larger sample of qass , particularly qass with large silicate optical depths , are needed to more strongly constrain the slope of the reddening - silicate optical depth relationship , and to investigate whether the relationship is adequately described by a simple linear trend . furthermore , measurements of the silicate optical depth in qass with low reddening values , in which the silicate optical depth is expected to be low requiring ( future ) high - sensitivity instruments , will help to establish whether the trend extends to systems with less dust . the 21-cm peak optical depth provides an estimate of the absorption produced by the cold , dusty h i gas in the ism . we list measurements of the observed @xmath110 , taken from the literature , in table [ compare ] . in the txs 0218 + 357 system , @xcite find a peak optical depth of @xmath111 , assuming a covering factor of 1.0 . the line structure is noted to be somewhat asymmetrical , but under an assumed gaussian fit a fwhm of 43@xmath44 km / s is derived . the velocity width , optical depth , and shape for this 21-cm absorption feature are all noted by @xcite to be similar to those seen in the z=0.437 qas toward 3c 196 , although concerns that the 3c 196 absorption is occurring against the lobes rather than against the agn core imply that they may not be directly comparable @xcite . this sightline difference may explain why despite the similarity in their measured 21-cm peak optical depths , the measured peak silicate optical depths differ : @xmath3 is nearly five times larger for the txs 0218 + 357 qas than for the z=0.437 qas toward 3c 196 . the remaining systems with 21 cm absorption measurements suggest that there may be a slight anti - correlation between the measured silicate peak optical depth and the measured 21 cm peak optical depth , as depicted in figure [ t21plot ] ( left ) . however , as evidenced by the poor match of the linear least - squares fit to the data , this may not be a purely linear relationship , but more data points are required to establish the precise shape of the relationship . when the covering factors are applied to the 21 cm and silicate peak optical depth measurements ( figure [ t21plot ] , left ) , the trend becomes less clear , largely because of the high derived @xmath83 optical depth for the pks 1830 - 211 system , which has a low covering factor . although a slight anti - correlation is suggested by the data , a larger data sample is required to investigate further . in any case , variations in properties of the material , such as the dust - to - gas ratio and the depletions , could induce scatter in even an intrinsically tight correlation . we also investigate suggestions of a correlation between the mg ii @xmath1122796 rest - frame equivalent width and the peak optical depth of the silicate dust absorption . @xcite found a correlation between these properties and suggested that the more silicate - rich qass may be more massive . this is because since the mg ii lines are saturated in the absorbers studied by @xcite , the mg ii equivalent width is a better proxy for the velocity spread ( i.e. the absorber mass or presence of outflows into the cgm ) than for the elemental abundance of metals in the ism gas . we have revisited the correlation here with the addition of the txs 0218 + 357 qas , as shown in figure [ mgplot ] using both the measured ( left ) and corrected ( right ) silicate peak optical depths . however , we find that the txs 0218 + 357 qas is an outlier relative to the other 5 qass , with a much stronger peak optical depth than would be predicted from its mg ii equivalent width . including all 6 qass and using the measured peak optical depths , these parameters have a pearson correlation coefficient of 0.22 ( 68% probability of no correlation ) , a spearman correlation coefficient of 0.71 ( significance 0.11 ) , and a kendall s tau value of 0.47 ( significance 0.19 ) . if we exclude the txs 0218 + 357 outlier point , however , we would obtain a stronger pearson correlation coefficient of 0.86 ( 6% probability of no correlation ) and kendall s tau of 0.60 ( significance 0.14 ) . a linear least - squares fit to all 6 qass , weighting solely by the @xmath3 uncertainties , since measurement uncertainties are not available for all of the mg ii measurements in the literature , produces a slope of [email protected] for the measured optical depth values , and [email protected] using the covering - factor corrected optical depths . this relationship is illustrated as a dashed line . the previously studied 5 qass cluster around this line , while the txs 0218 + 357 qas lies substantially above the fit . if the txs 0218 + 357 qas is excluded from the fit , the slope decreases slightly to [email protected] ( same for the corrected optical depths ) , as depicted with the dotted line . a larger sample will be required to see whether the txs 0218 + 357 qas is anomalous , or whether the suggested correlation is not as tight was was implied with the earlier sample of qass . some of the discrepancy for the txs 0218 + 357 qas may arise if the mg ii line in this system is not saturated , unlike in the 5 previously studied qass . this may be caused in part by the mg ii slit encompassing both quasar sightlines , one of which has significantly less absorption and is brighter at optical wavelengths . while the equivalent width ratio of the lines is suggestive of saturation , the resolution and s / n of the spectrum are not high enough to definitively assess if the lines are saturated @xcite . if the lines are unsaturated , the mg ii measurement may be providing an estimate more reflective of the elemental abundance than of the galaxy dynamics ( i.e. the velocity spread ) . alternatively , it is possible that the sampled sightline through the txs 0218 + 357 qas s nearly face - on galaxy probes inner regions which produce less efficient galactic outflows of warm gas containing mg ii , and that if viewed from a different sightline a higher mg ii equivalent width would be obtained . if an elemental discrepancy is invoked , the ism gas in this system may be mg - poor either because of an intrinsic deficit in the total ( gas+solid phase ) amount of mg , which would also result in less mg - rich ( more fe - rich ) species of olivines / pyroxene silicates . alternatively , the gas may be mg - poor because of depletions onto the dust grains , resulting in mg - rich olivines / pyroxenes . data with higher spectral resolution and s / n , and with longer wavelength coverage , are required to discriminate better between these silicate species . lastly , the txs 0218 + 357 qas provides an interesting opportunity to investigate whether there is a link between silicate crystallinity and a high enrichment of molecular material . the degree of silicate crystallinity depends on the conditions ( e.g. , temperature ) under which the dust grains formed and the subsequent processing of the dust grains ( e.g. , by cosmic rays , grain - grain collisions ) , and , therefore , has implications for the evolution of the dust grains [ e.g. , @xcite ] . in the z=0.886 qas toward pks 1830 - 211 , a system which has been noted to have the largest number ( 28 ) of detected molecular species in any extragalactic object to date @xcite , we previously found very suggestive evidence for a high degree of silicate crystallinity @xcite . this evidence for crystallinity stems from the shape and location of the substructure in the observed qas 10 @xmath1 silicate absorption feature . since in the milky way ism the silicate dust is predominately amorphous @xcite , this raises the question of how the silicate dust in pks 1830 - 211 could be so much more crystalline , and whether it is linked to the high degree of molecule - enrichment . for instance , does some aspect ( e.g. temperature , density ) of the ism environment facilitate the formation of both molecules and crystalline silicate grains ? since the txs 0218 + 357 qas is also noted to be very rich in molecules , and exhibits 10 @xmath1 silicate absorption , it provides a unique opportunity to test this hypothesis . the txs 0218 + 357 qas has been noted to exhibit signatures of many molecules , including co , hco@xmath28 , hcn , hnc , cn , cs , c@xmath29o , @xmath30co , h@xmath8o , nh@xmath31 , oh , h@xmath8co , and lih @xcite . as for the lensed pks 1830 - 211 z=0.886 qas , it appears that one of the two quasar sightlines ( both of which are encompassed by the _ irs _ slit ) is more enriched . although a direct comparison of the abundances in the z=0.886 pks 1830 - 211 qas and the txs 0218 + 357 qas is challenging , since derived molecular abundances depend on assumptions about the degree of line saturation , the covering factor , and the excitation temperature , for most molecules it appears that the txs 0218 + 357 qas is slightly less enriched . the oh and hco@xmath28 column densities are both good predictors of n@xmath113 in galactic clouds @xcite . a direct comparison in @xcite finds that the derived n@xmath114 and n@xmath115 column densities in txs 0218 + 357 are 19 - 23% of those for the z=0.886 pks 1830 - 211 absorber . using oh as a tracer of the molecular hydrogen results in a column density of n@xmath113 of @xmath116 @xmath12 for the txs 0218 + 357 @xcite . however , while the evidence for significant molecule enrichment in the ism in the txs 0218 + 357 qas is undisputed , the evidence for or against silicate crystallinity is much less clear due to the low s / n of our data . the combination of a lower flux density , fewer exposures , and a smaller total exposure time for the txs 0218 + 357 quasar , in comparison with pks 1830 - 211 @xcite , makes it much more difficult to establish the presence or absence of silicate crystallinity . as discussed in [ results ] the laboratory amorphous olivine laboratory template for the txs 0218 + 357 system provides a good fit , both in terms of the breadth and depth of the 10 and 18 @xmath1 silicate absorption features , and in terms of the peak absorption wavelength . however , as illustrated in figure [ singlefit ] , the magnitude of the substructure in the 10 and 18 @xmath1 regions for some species of crystalline olivine and pyroxene are small enough to be masked by the large uncertainties and noise associated with our spectrum . several of the predicted substucture features align relatively well with the possibly noise features present in our data . furthermore , as established with the two - mineral fits in [ twofitsect ] , a combination of crystalline olivine and pyroxene would produce an excellent match to the 10 @xmath1 and 18 @xmath1 silicate absorption features ; the predicted abundant substructure would occur at levels below that of the noise in the data . thus , although the smooth , featureless amorphous olivine fit is both consistent with the ism seen in the milky way and with our data , we can not rule out that abundant silicate crystallinity _ could _ be present in the txs 0218 + 357 , although it seems less likely than in the pks 1830 - 211 z=0.886 qas . if the explanation that the txs 0218 + 357 qas is produced by amorphous olivines is adopted , it establishes a similarity between the silicate dust in this z=0.685 galaxy and that in the milky way ism . by contrast , independent of the substructure , the redder peak wavelength of the absorption feature in the pks 1830 - 211 z=0.886 qas implies silicate dust grain differences with that in the milky way ism and with that in the txs 0218 + 357 qas ism . in combination , these data suggest that there could be significant variations in the dust grain properties between different galaxies , which may not be a simple function of redshift evolution or of molecular gas abundance . further investigation of the silicate dust grains in the txs 0218 + 357 with higher spectral resolution and s / n data , as well as a study of a broader sample of qass , is required to quantify these variations and establish whether ism silicate dust grain crystallinity varies between galaxies as a function of composition , star - formation rate , or redshift . such studies may be possible within the next decade using new facilities such as _ sofia _ and _ jwst - miri_. in summary , we find clear evidence of 10 @xmath1 and 18 @xmath1 silicate absorption in the txs 0218 + 357 qas . this shape and peak wavelength of these absorption features appears to be best represented by an amorphous olivine template , with a measured peak optical depth of @xmath117 , although we can not rule out the possibility of substructure originating in crystalline silicates in the relatively poor s / n data . a quantitative assessment of the silicate crystallinity in this system will require higher s / n data and detection of longer - wavelength mid - ir crystalline resonance features . fits using amorphous pyroxene and sic templates poorly matched the observed silicate absorption profile for the txs 0218 + 357 qas . the derived optical depth is significantly higher than that found for other qass , although the rest - frame mg ii equivalent width and the optical depth of the 21-cm absorption are within the range of values observed for the other program qass . this may suggest that the silicate dust in the txs 0218 + 357 system is not accompanied by a substantial enhancement in gas . we also find that while the reddening in the txs 0218 + 357 system is higher than in other qass , it is not as high as the large optical depth would predict , given an extrapolation between these two parameters based on milky way diffuse clouds . this may suggest variations in the silicate dust grain properties between different galaxies , which can be investigated further with larger samples . we thank the anonymous referee for his / her comments , and dr . eli dwek ( nasa - goddard ) for conversations which improved the content of this paper . this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration ( nasa ) . this work is based on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa . support for this work was provided by nasa through an award issued by jpl / caltech . mca and vpk also gratefully acknowledge partial support from the nsf grant ast-1108830 ( pi kulkarni ) to the university of south carolina . _ facilities : _ in order to determine the effect on our conclusions of variations in the adopted underlying quasar continuum shape , we explore in this appendix alternative quasar continuum normalizations . the third order chebyshev polynomial fit we adopt in our analysis was selected because it visually provided the best match to the quasar continuum shape . we also explored 22 other fits to the continuum , produced by linear , power law , third order chebyshev polynomial , and cubic spline fits . we found that all of the fits were similar near 10 @xmath1 , while there were variations in the more poorly constrained 18 @xmath1 region because of the paucity of data at @xmath11821 @xmath1 . we summarize in table [ contin ] the two most dissimilar fits over the silicate features . lllll fiducial & 3rd order chebyshev & 8.5 - 12.5 ; 16.5 - 20.4 & ll1/ll2 join & adopted in main body of paper + altnorm1 & 3rd order chebyshev & 8.5 - 12.5 ; 16.5 - 20.4 & ll1/ll2 join ; @xmath112 near 21 @xmath1 & shallower 18 @xmath1 feature + altnorm2 & linear & 8.5 - 12.5 ; 16.5 - 20.4 & & deeper 18 @xmath1 feature + [ contin ] the two quasar continuum normalizations with the largest differences in the shapes of the silicate absorption features were obtained using the chebyshev polynomial and linear continuum fits illustrated in figure [ norm - fig3 ] . the first normalization ( altnorm1 ) uses a third order chebyshev polynomial like the fiducial fit , but rejects several additional points at @xmath11920 @xmath1 which produces a shallower and slightly narrower 18 @xmath1 absorption feature . the second normalization ( altnorm2 ) is a weighted linear fit , excluding the regions associated with the 10 @xmath1 silicate feature ( 8.5 - 12.5 @xmath1 ) and the 18 @xmath1 feature ( 16.5 - 20.4 @xmath1 ) , and results in a slightly deeper and broader 18 @xmath1 feature . the differences in the 10 @xmath1 region are minimal for both normalizations . the shallower 18 @xmath1 feature would have a rest - frame equivalent width of [email protected] ( a 3.3@xmath2 detection ) , while the the broader feature would have a larger rest - frame equivalent - width of [email protected] ( a 4.4@xmath2 detection ) . as previously discussed , if the feature extends redward of our data , these are lower limits . for these 2 alternative quasar continuum normalizations , we repeated the fitting described in [ results ] . we did not alter the @xmath48 minimization region in the fitting , although the breadth of the features vary slightly between the continuum normalizations . we found that when fitting over the 10 @xmath1 region alone the alternative quasar continuum normalizations had a small impact on the derived peak optical depth . for both normalizations we again obtained that the amorph.oliv.gpc template produced the best fit . the measured peak optical depth for this fit was @xmath120 for the altnorm1 normalization and @xmath121 for the altnorm2 normalization . while these are both slightly lower than the measured @xmath68 peak optical depth obtained using the fiducial normalization , they are consistent within 1.5@xmath2 . we illustrate in figure [ norm - fig4 ] these fits over only the 10 @xmath1 feature in the left two panels , with the am.oliv.gpc template fit shown in red . when we extended the fitting region to cover both the 10 @xmath1 and 18 @xmath1 features , the differences in the quasar continuum normalization were more significant . the shallower 18 @xmath1 feature in the altnorm1 normalization results in a 10:18 @xmath1 optical depth ratio that is more consistent with considered crystalline pyroxene templates than with those for amorphous olivine . furthermore , using the best - fitting minerals , the derived peak optical depth for the 10 @xmath1 feature was found to be deeper than derived for the fiducial fit : @xmath122 versus the fiducial normalization fit of @xmath69 , as shown in table [ fits-10 + 18-apb ] . we illustrate this crystalline pyroxene fit in cyan in figure [ norm - fig4 ] . while more substructure is associated with the crystalline pyroxene fit , it is not inconsistent with our data given the relatively low s / n . for the altnorm2 normalization the peak optical depth and best - fitting mineral were consistent with the fiducial normalization . lllllll amorph . olivine & amorph.oliv.gpc ( 1 ) & [email protected] & 0.91 & 79.9 & [email protected] & 0.70 + amorph . pyroxene & amorph.pyrox.gpc ( 1 ) & [email protected]&1.08&21.6&[email protected] & 1.34 + cryst . olivine & sancarl.oliv.(2 ) & [email protected]&1.15&8.8&[email protected] & 1.31 + cryst . pyroxene ( best ) & nat.enstatite ( 3 ) & 0.57@xmath123&0.88&87.2&[email protected] & 1.00 + amorph . olivine ( best ) & amorph.oliv.gpc ( 1 ) & [email protected]&0.71&99.9&[email protected] & 0.70 + amorph . pyroxene & amorph.pyrox.gpc ( 1 ) & [email protected]&1.80&0.0&[email protected] & 1.34 + cryst . olivine & olivine ( mg@xmath124fe@xmath62sio@xmath51 ) ( 3 ) & [email protected]&1.16&7.1 & [email protected] & 1.27 + cryst . pyroxene & nat.enstatite ( 3 ) & [email protected]&1.08&20.8&[email protected] & 1.00 + [ fits-10 + 18-apb ] in order to determine the optimal peak optical depth and place _ some _ constraints on the mineralogy of the silicate dust grains producing the observed 10 and 18 @xmath1 absorption features , we considered a range of absorption profile templates from the literature . the full list of considered minerals is presented in table 10 of @xcite , where we discuss the methodology for selecting these templates . since our goal in this analysis is to identify plausible mineral groups , not to identify the precise mineral grain properties , we discuss only the minerals which best - fit the txs 0218 + 357 qas silicate features in each of the mineral categories : amorphous olivines and pyroxenes , crystalline olivines and pyroxenes , and sic sources . we follow conventions in the literature when discussing ` amorphous ' silicates . we note , however , that amorphous olivine / pyroxene physically refers to amorphous silicates of an olivine / pyroxene composition , since all olivines / pyroxenes are intrinsically crystalline @xcite . furthermore , we recognize that the nature of many of the amorphous olivine / pyroxene samples in the literature have been called into question , either because of the condensation methods which may impart impurities or because of the existence of crystalline or semi - crystalline grains in the putatively amorphous samples @xcite . these effects do not significantly impact our conclusions . aller , m. c. , kulkarni , v. p. , york , d. g. , et al . 2012 , , 748 , 19 bernstein , r. a. , freedman , w. l. , & madore , b.f . 2002 , , 571 , 107 bowey , j. e. , & adamson , a. j. 2001 , , 320 , 131 briggs , f. h. , de bruyn , a. g. , & vermeulen , r. c. 2001 , , 373 , 113 browne , i. w. a. , patnaik , a. r. , walsh , d. , & wilkinson , p. n. 1993 , , 263 , l32 carilli , c. l. , rupen , m. p. , & yanny , b. 1993 , , 412 , l59 chengalur , j. n. , de bruyn , a. g. , & narasimha , d. 1999 , , 343 , l79 chiar , j. e. , ennico , k. , pendleton , y. j. , et al . 2007 , , 666 , l73 chiar , j. e. , & tielens , a. g. g. m. 2006 , , 637 , 774 cohen , j. g. , lawrence , c. r. , & blandford , r. d. 2003 , , 583 , 67 combes , f. , & wiklind , t. 1995 , , 303 , l61 combes , f. , & wiklind , t. 1997 , , 486 , l79 combes , f. , & wiklind , t. 1998 , , 334 , l81 combes , f. , wiklind , t. , & nakai , n. 1997 , , 327 , l17 curran , s. j. , murphy , m. t. , pihlstrm , y. m. , webb , j. k. , & purcell , c. r. 2005 , , 356 , 1509 dorschner , j. , begemann , b. , henning , t. , jger , c. , & mutschke , h. 1995 , , 300 , 503 draine , b. t. 2003 , , 41 , 241 dwek , e. , & cherchneff , i. 2011 , , 727 , 63 fabian , d. , henning , t. , jger , c. , et al . 2001 , , 378 , 228 falco , e. e. , impey , c. d. , kochanek , c. s. , et al . 1999 , , 523 , 617 foltz , c. b. , chaffee , f. h. , jr . , & wolfe , a. m. 1988 , , 335 , 35 forrest , w. j. , gillett , f. c. , & stein , w. a. 1975 , , 195 , 423 friedel , d. n. , kemball , a. , & fields , b. d. 2011 , , 738 , 37 friedemann , c. , guertler , j. , schmidt , r. , & dorschner , j. 1981 , , 79 , 405 frye , b. , welch , w. j. , & broadhurst , t. 1997 , , 478 , l25 gerin , m. , phillips , t. g. , benford , d. j. , et al . 1997 , , 488 , l31 gharanfoli , s. , kulkarni , v. p. , chun , m. r. , & takamiya , m. 2007 , , 133 , 130 greenberg , j. m. , & li , a. 1999 , adv . space res . , 24 , 497 hao , l. , spoon , h. w. w. , sloan , g. c. , et al . 2005 , , 625 , l75 henkel , c. , braatz , j. a. , menten , k. m. , & ott , j. 2008 , , 485 , 451 henkel , c. , jethava , n. , kraus , a. , et al . 2005 , , 440 , 893 henning , th . 2010 , , 48 , 21 henning , th . , ilin , v. b. , krivova , n. a. , michel , b. , & voshchinnikov , n. v. 1999 , , 136 , 405 houck , j. r. , roellig , t. l. , van cleve , j. , et al . 2004 , , 154 , 18 jger , c. , molster , f. j. , dorschner , j. , et al . 1998 , , 339 , 904 jethava , n. , henkel , c. , menten , k. m. , carilli , c. l. , & reid , m. j. 2007 , , 472 , 435 jones , a. p. , & nuth , j. a. , iii 2011 , , 530 a44 jones , a. p. , tielens , a. g. g. m. , & hollenbach , d. j. 1996 , , 469 , 740 junkkarinen , v. t. , cohen , r. d. , beaver , e. a. , et al . 2004 , , 614 , 658 kanekar , n. , & chengalur , j. n. 2002 , , 381 , l73 kanekar , n. , chengalur , j. n. , de bruyn , a. g. , & narasimha , p. 2003 , , 345 , l7 kanekar , n. , lane , w. m. , momjian , e. , briggs , f. h. , & chengalur , j. n. 2009 , , 394 , l61 kemper , f. , markwick , a. j. , & woods , p. m. 2011 , , 413 , 1192 kemper , f. , vriend , w. j. , & tielens , a. g. g. m. 2004 , , 609 , 826 khare , p. , vanden berk , d. , york , d. g. , lundgren , b. , & kulkarni , v. p. 2012 , , 419 , 1028 koike , c. , mutschke , h. , suto , h. , et al . 2006 , , 449 , 583 kulkarni , v. p. , torres - garcia , l. m. , som , d. , et al . 2011 , , 726 , 14 kulkarni , v. p. , york , d. g. , vladilo , g. , & welty , d. e. 2007 , , 663 , l81 lehr , j. , falco , e. e. , kochanek , c. s. , et al . 2000 , , 536 , 584 li , a. , & draine , b. t. 2001 , , 554 , 778 li , m. p. , zhao , g. , & li , a. 2007 , , 382 , l26 liszt , h. , & lucas , r. 1999 , in highly redshifted radio lines " , ed . c. l. carilli , s. j. e. radford , k. m. menten , & g. i. langston , asp conf . ser . , 156 , 188 menten , k. m. , & reid , m. j. 1996 , , 465 , l99 messenger , s. j. , speck , a. , & volk , k. 2013 , , 764 , 142 molster , f. , & kemper , c. 2005 , , 119 , 3 muller , s. , beelen , a. , gulin , m. , et al . 2011 , , 535 , a103 muller , s. , gulin , m. , combes , f. , & wiklind , t. 2007 , , 468 , l53 odea , c. p. , baum , s. a. , stanghellini , c. , et al . 1992 , , 104 , 1320 patnaik , a. r. , browne , i. w. a. , king , l. j. , et al . 1993 , , 261 , 435 patnaik , a. r. , browne , i. w. a. , wilkinson , p. n. , & wrobel , j. m. 1992 , , 254 , 655 patnaik , a. r. , porcas , r. w. , & browne , i. w. a. 1995 , , 274 , l5 pei , y. c. , fall , s. m. , & bechtold , j. 1991 , , 378 , 6 proux , c. , dessauges - zavadsky , m. , dodorico , s. , kim , t .- s . , & mcmahon , r. g. 2005 , , 363 , 479 pettini , m. , king , d. l. , smith , l. j. , & hunstead , r. w. 1997 , , 478 , 536 prochaska , j. x. , herbert - fort , s. , & wolfe , a. m. 2005 , , 635 , 123 roberts , m. s. , brown , r. l. , brundage , w. d. , et al . 1976 , , 81 , 293 roche , p. f. , & aitken , d. k. 1984 , , 208 , 481 roche , p. f. , & aitken , d. k. 1985 , , 215 , 425 sembach , k. r. , & savage , b. d. 1992 , , 83 , 147 speck , a. k. , corman , a. b. , wakeman , k. , wheeler , c. h. , & thompson , g. 2009 , , 691 , 1202 speck , a. k. , whittington , a. g. , & hofmeister , a. m. 2011 , , 740 , 93 spoon , h. w. w. , tielens , a. g. g. m. , armus , l. , et al . 2006 , , 638 , 759 srianand , r. , gupta , n. , petitjean , p. , noterdaeme , p. , & saikia , d. j. 2008 , , 391 , l69 storrie - lombardi , l. j. & wolfe , a. m. 2000 , , 543 , 552 van breemen , j. m. , min . , m. , chair , j. e. , et al . 2011 , , 526 , a152 weingartner , j. c. , & draine , b. t. 2001 , , 548 , 296 wiklind , t. , & combes , f. 1995 , , 299 , 382 wolfe , a. m. , & wills , b. j. 1977 , , 218 , 39 york , d. g. , khare , p. , vanden berk , d. , et al . 2006 , , 367 , 945 york , t. , jackson , n. , browne , i. w. a. , wucknitz , o. , & skelton , j. e. 2005 , , 357 , 124 zeiger , b. , & darling , j. 2010 , , 709 , 386 zubko v. , dwek , e. , & arendt , r.g . 2004 , , 152 , 211
we report the detection of interstellar silicate dust in the z@xmath0=0.685 absorber along the sightline toward the gravitationally lensed blazar txs 0218 + 357 . using _ spitzer space telescope _ _ infrared spectrograph _ data we detect the 10 @xmath1 silicate absorption feature with a detection significance of 10.7@xmath2 . we fit laboratory - derived silicate dust profile templates obtained from literature to the observed 10 @xmath1 absorption feature , and find that the best single - mineral fit is obtained using an amorphous olivine template with a measured peak optical depth of @[email protected] , which rises to @[email protected] if the covering factor is taken into account . we also detected the 18 @xmath1 silicate absorption feature in our data with a @xmath63@xmath2 significance . due to the proximity of the 18 @xmath1 absorption feature to the edge of our covered spectral range , and associated uncertainty about the shape of the quasar continuum normalization near 18 @xmath1 , we do not independently fit this feature . we find , however , that the shape and depth of the 18 @xmath1 silicate absorption are well - matched to the amorphous olivine template prediction , given the optical depth inferred for the 10 @xmath1 feature . the measured 10 @xmath1 peak optical depth in this absorber is significantly higher than those found in previously studied quasar absorption systems . the reddening , 21-cm absorption , and velocity spread of mg ii are not outliers relative to other studied absorption systems , however . this high optical depth may be evidence for variations in dust grain properties in the ism between this and the previously studied high redshift galaxies .
the problem of finding global extrema ( maxima and minima ) for a univariate real function is important for variety of real world applications . for example , it arises in electric engineering @xcite,@xcite , @xcite in computer science @xcite , @xcite , @xcite and in various other fields ( see @xcite for further references ) . in many industrial applications the global optimization algorithm is expected to operate in real time while simultaneously , finding the global extremum of non - convex functions exhibiting large number of local sub - extrema . the problem of efficiently finding a function s global extremum has been historically challenging . one of the first solutions , zero derivative method ( zdm ) , was proposed by pierre de fermat ( 1601 - 1665 ) . his main idea was to look for the global extremum among critical points : the points where the derivative of the target function is zero . despite its theoretical significance , fermat s proposed method ( zdm ) is limited by the numerical difficulties imposed by finding critical points . one of the leading global optimization approaches adopted by many industrial applications is a brute - force search or exhaustive search for the global extremum ( brute - force search ( bfs ) ) . it is simple to implement but its performance linearly depends on the complexity of the target function and the size of the search area . a plethora of optimization methods have been developed for various types of target functions . among them piyavskii - shubert method ( psm ) occupies a special place @xcite , @xcite , @xcite . it is one of the few procedures that delivers the global extremum for a univariate function and at the same time it exhibits reasonable performance as long as the respective lipschitz constant is of a modest value . on the other hand , the method is very sensitive to the size of the lipschitz constant : its performance sharply diminishes for large lipschitz constants . for this reason , accelerations and improvements of psm were developed in the following papers @xcite , @xcite , @xcite , @xcite . numerical experiments presented in this paper show that leap gradient algorithm ( referred as lga ) significantly outperforms psm together with its modifications and improvements ( from @xcite , @xcite , @xcite ) when finding global extrema of polynomials . the method of gradient descent ( see , e.g. @xcite , @xcite , @xcite ) is widely used to solve various practical optimization problems . the main advantage of the gradient descent algorithm is its simplicity and applicability to a wide range of practical problems . on the other hand , gradient descent has limitations imposed by the initial guess of a starting point and then its subsequent conversion to a suboptimal solution . this paper gives practical recipes on how to overcome those limitations for a univariate function and how to equip the gradient descent algorithm with abilities to converge to a global extremum . it is achieved via evolutionary leaps towards the global extremum . lga neither requires the knowledge of the lipschitz constant nor convexity conditions that are often imposed on the target function . moreover , lga naturally becomes the standard gradient descent procedure when the target function is convex or when the algorithm operates in the close proximity to the global extremum . the recursive application of lga yields an efficient algorithm for calculating global extrema for univariate polynomials . lga does not intend to locate any critical points ( like zdm ) instead it follows gradient descent ( ascent ) until a local extremum is reached and then performs an evolutionary leap towards the next extremum . as far as performance is concerned , numerical experiments conducted for univariate polynomials show that lga outperforms bfs , zdm and psm with all its modifications from @xcite , @xcite , @xcite . the layout of this publication is as follows . section [ intro ] is the introduction . section [ sec:2 ] introduces lga for univariate functions . section [ lgarecursion ] explores the recursive application of lga . section [ multiw ] describes a recursive implementation of lga for polynomials . section [ realanalytic ] presets an implementation of lga for real analytic univariate functions . section [ numerical_experiments ] reports on numerical experiments with lga . it compares lga with zdm , bfs , psm and accelerations of psm presented in @xcite , @xcite , @xcite . polynomial roots for zdm are calculated with the help of laguerre s method @xcite , @xcite . section [ acknowledgment ] expresses gratitude to professionals who spent their time contributing to the paper , reviewing various parts of the publication and making many valuable suggestions and improvements . section [ appendix ] finalizes the publication with a snapshot of the working and thoroughly tested source code for lga . let @xmath0 be known and well defined real function on @xmath1,$ ] a closed interval of real numbers . consider the optimization problem @xmath2}\ ] ] lga can solve it with a given precision @xmath3 if its solution exists . [ [ leap - gradient - algorithm - lga . ] ] * leap gradient algorithm ( lga ) . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + * set @xmath4 * iterate @xmath5 as long as @xmath6 and @xmath7 if @xmath8 then * stop * and take @xmath9 as an estimate of the argument and the value for the global minimum . + if @xmath10 then proceed with * step 2*. * let @xmath11 be the solution of the following optimization problem @xmath12}\ ] ] if @xmath13 then * stop * and @xmath14 is an estimate of the argument and the value for the global minimum . + if @xmath15 then @xmath16 ( lga performs _ an evolutionary leap _ ) and go to * step 1*. lga is illustrated in fig . [ fig:1 ] . epsf if the target function is twice continuously differentiable on @xmath17 then the number of inflection points from @xmath17 is related to the number of evolutionary leaps performed by lga ( see * step 2 * ) . [ inflection_and_leaps ] let @xmath0 be twice continuously differentiable on @xmath17 . let @xmath18 be the value of the evolutionary leap ( * step 2 * of lga ) . if @xmath19 then @xmath20 $ ] contains at least two points where @xmath21 is equal to zero . evolutionary leaps from @xmath22 ( on @xmath23 are solutions of the following equation . @xmath24 where @xmath25 indeed , if @xmath0 is continuously differentiable then the necessary condition for @xmath26 to be the solution for ( [ optimization_proc ] ) on @xmath27 is @xmath28 differentiating yields @xmath29 after multiplying the equation with @xmath30 we obtain ( [ leap_formula ] ) . since @xmath18 is the solution of the problem ( [ optimization_proc ] ) and @xmath0 is twice differentiable we conclude that @xmath31 differentiating yields @xmath32 when @xmath33 hence , making use of ( [ leap_formula ] ) we obtain @xmath34 according to * step 2 * of lga for the evolutionary leap @xmath18 we have @xmath35 that together with ( [ leap_formula ] ) implies @xmath36 on the other hand , ( [ gradient_decent ] ) yields the existence of @xmath37 such that @xmath38 that together with ( [ rt ] ) yield the existence of @xmath39 where @xmath40 taking into account that @xmath41 is obtained after * step 1 * of lga and @xmath42 we obtain the existence of @xmath43 where @xmath44 under the conditions of the theorem @xmath45 is continuous on @xmath17 and the statement follows from ( [ rightplus ] ) , ( [ midminus ] ) and ( [ leftplus ] ) . that means ( in a generic situation ) a non trivial evolutionary leap inside @xmath27 is only possible over two inflection points of a target function . consider the optimization problem @xmath46 where @xmath47 and @xmath48 are real numbers . if @xmath49 then @xmath50 has two different zeroes . hence , by theorem [ inflection_and_leaps ] , lga might need to perform a single evolutionary leap ( over two inflection points ) in order to solve the optimization problem . otherwise , @xmath51 , lga coincides with the standard gradient decent . lga replaces the optimization problem @xmath52}\ ] ] with @xmath53}\ ] ] where @xmath22 is calculated at the previous step of lga . it leads us to the following recursive procedure executed at each iteration of lga . @xmath54 and @xmath55 in order to complete an iteration of lga for @xmath56 one needs to calculate @xmath57}\ ] ] and lga finds the minimum of @xmath58 if finding the minimum for @xmath59 is obvious , then the iteration of lga is completed after @xmath60 recursive steps . the next theorem shows when an lga iteration is completed after a finite number of recursive steps . if a real function @xmath0 is at least @xmath61-times continuously differentiable on the segment @xmath1 $ ] and @xmath62\ ] ] then each iteration of lga for @xmath52}\ ] ] is completed after not more than @xmath63 recursive steps . the function can be represented by taylor expansion centered at @xmath64 @xmath65 for all @xmath66.$ ] in accordance with notations ( [ iter_1 ] ) and ( [ iter_2 ] ) we have @xmath67 and @xmath68 under the conditions of the theorem @xmath69 achieves its minimum on @xmath70 $ ] at @xmath71 recursive lga is an efficient numerical method for finding global extrema of univariate polynomials . for any polynomial @xmath72 on a segment @xmath1 $ ] recursive lga delivers the global extremum for @xmath73 in a finite number of steps . the proof is conducted by mathematical induction with respect to the degree of a polynomial . as a basis of mathematical induction and for the purpose of illustrations let us consider finding the global minimum on @xmath1 $ ] for the quadratic polynomial @xmath74 in accordance with notations ( [ iter_1 ] ) , ( [ iter_2 ] ) @xmath75 if @xmath76 then @xmath77 has its minimum at @xmath78 if @xmath79 then according to lga the global minimum is reached ad @xmath80 and the procedure stops . otherwise , @xmath81 lga leads us to @xmath82 where the step size @xmath83 is dictated by the required precision of lga . we follow the standard gradient descent when calculating @xmath84 as long as @xmath85 the gradient descent either stops at @xmath86 and the global minimum is located at @xmath86 or , according to lga , it stops at the first @xmath22 such that @xmath87 and @xmath22 delivers the global minimum . if @xmath88 then the minimum for @xmath77 is located at @xmath89 according to lga , if @xmath90 then the minimum is at @xmath91 otherwise , the minimum is at @xmath89 the basis of the mathematical induction is established . the step of mathematical induction follows directly from recursive lga . indeed , suppose that recursive lga delivers , in a finite number of steps , the global minimum for any polynomial of degree less than @xmath92 however , following notations ( [ iter_1 ] ) and ( [ iter_2 ] ) , in order to calculate the global minimum for a polynomial @xmath93 of @xmath61-th degree we need to calculate the global minimum for @xmath94 a polynomial of @xmath95-st degree . hence , the statement follows by mathematical induction . horner s algorithm ( see , e.g. , @xcite , @xcite ) plays the central role in implementation of recursive lga for polynomials . lga reduces the optimization problem @xmath52}\ ] ] to @xmath96}\ ] ] where @xmath22 is calculated at the previous step of lga . if @xmath97 is a polynomial with real coefficients then so is @xmath98 where coefficients @xmath99 are calculated as follows . [ [ horners - algorithm ] ] * horner s algorithm * + + + + + + + + + + + + + + + + + + + + + * @xmath100 * @xmath101 recursive lga ( [ iter_1 ] ) , ( [ iter_2 ] ) described in section [ lgarecursion ] is reduced to a finite number of iterations for horner s algorithm until the resulted polynomial is either a linear or a quadratic function . then the solution of the optimization problem is trivial and therefore the recursive procedure delivers the global extremum . performing an evolutionary leap is a numerically expensive operation . theorem [ inflection_and_leaps ] shows that each lga leap inside the interval is a jump over two zeroes of the second derivative of the target function . that allows to improve the performance of lga for polynomials by limiting the number of evolutionary jumps by at most @xmath102 where @xmath61 is the degree of the target polynomial . the respective modification of lga for @xmath103}\ ] ] is as follows . [ [ lga - for - polynomials ] ] * lga for polynomials * + + + + + + + + + + + + + + + + + + + + + * if the degree of the polynomial is @xmath104 then return @xmath105 otherwise , return @xmath106 as the argument , value pair for the global minimum . + if the degree is equal to @xmath107 @xmath108 if @xmath109 then return @xmath110 for @xmath111 and return @xmath112 when @xmath113 otherwise , return @xmath114 where @xmath115 + if the degree of the polynomial is larger than @xmath107 then set @xmath116 @xmath4 * iterate @xmath5 as long as @xmath117 and @xmath118 where @xmath119 are calculated with horner s algorithm . * * @xmath120 * * @xmath121 + if @xmath8 then * stop * and return @xmath86 as the argument and @xmath122 as the value for the estimate of the global minimum . + if @xmath123 then proceed with * step 2*. * if lga already performed @xmath124 evolutionary jumps , @xmath125 then * stop * and return @xmath22 and @xmath126 as the argument , value pair for the estimate of the global minimum . + otherwise , recursively apply lga to @xmath127}\ ] ] + let @xmath11 be the solution of ( [ recursive_lga ] ) . if @xmath128 and @xmath129 then * stop * and return @xmath64 @xmath126 as an argument , value estimate of the global minimum . + if @xmath130 and @xmath131 then , by theorem [ inflection_and_leaps ] , increment @xmath132 by one if @xmath22 equals to @xmath133 otherwise by two . after that lga performs _ an evolutionary leap _ by setting @xmath134 and proceeding with * step 1*. a polynomial @xmath126 in lga is evaluated with horner s algorithm as follows . set @xmath135 then iterate @xmath136 and @xmath137 is the value of the polynomial at @xmath138 interested reader will find a snapshot of the working and thoroughly tested source code of lga in appendix of this paper . consider the optimization problem @xmath139}\ ] ] where @xmath0 is a univariate real analytic function on @xmath140.$ ] that means @xmath141 and the series uniformly converges to @xmath0 on @xmath140.$ ] @xmath142 denotes the value of the derivative @xmath143 at @xmath144 this section proposes a numerical procedure for solving ( [ optimization_problem_analytic ] ) . the procedure is based on lga . namely , ( [ optimization_problem_analytic ] ) is replaced with @xmath145}\ ] ] where @xmath146 and @xmath147 the numerical algorithm is based on the following statement . [ real_analytic_lga ] for any @xmath148 $ ] @xmath149 it follows from taylor expansion with the integral remainder term that @xmath150 in order to calculate @xmath151 let us justify the following statements with the help of mathematical induction . @xmath152 the basis of mathematical induction follows from @xmath153 the step of mathematical induction for each of the statements ( [ first ] ) , ( [ second ] ) , ( [ third ] ) is as follows . suppose that ( [ first ] ) is true for @xmath154 and @xmath155 to prove that it remains true for @xmath156 and @xmath157 consider @xmath158 and @xmath159 due to the assumption of the mathematical induction @xmath160 for @xmath161 the statement ( [ first ] ) follows . assume that ( [ second ] ) is true for @xmath162 consider @xmath163 and ( [ first ] ) together with the assumption of the mathematical induction yield @xmath164 the statement ( [ second ] ) is established . assume that ( [ third ] ) is valid for @xmath162 then @xmath165 by the assumption of the mathematical induction and taking into account ( [ first ] ) , ( [ second ] ) we have @xmath166 statement ( [ third ] ) is established . applying @xmath167 to the taylor expansion @xmath168 and taking into account ( [ first ] ) , ( [ second ] ) yields @xmath169 on the other hand , @xmath170 and so is @xmath171 therefore setting @xmath172 in ( [ fourth ] ) and making use of ( [ third ] ) we obtain @xmath173 that completes the calculation of ( [ difference ] ) and the proof . given the required margin of error @xmath174 theorem [ real_analytic_lga ] provides an effective recipe for finding the global minimum of a real analytic function on the interval @xmath140.$ ] [ [ lga - for - real - analytic - functions ] ] * lga for real analytic functions * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + * find a natural number @xmath61 so that @xmath175 * calculate a step size @xmath3 such that @xmath176 * use lga to solve optimization problem ( [ optimization_problem_lga ] ) with @xmath3 from * step 2*. let @xmath177 be the @xmath26-argument of the global minimum calculated at * step 3*. let @xmath178 be the @xmath26-argument of the global minimum calculated for the original function @xmath0 from ( [ optimization_problem_analytic ] ) . then @xmath179 and @xmath180 this section presents the results of numerical experiments conducted in order to compare the performance of lga with brute - force search ( bfs ) , zero derivative method ( zdm ) , modifications of piyavskii - shubert method ( psm ) discussed in @xcite , @xcite , @xcite . all numerical experiments follow the same scenario : * repeat 500 times * step 1 * and * step 2*. * * randomly generate a real polynomial @xmath73 of degree @xmath61 with roots uniformly distributed on @xmath181\times [ -1,\;1],$ ] where @xmath86 is a real number between @xmath182 and @xmath183 which remains fixed across all 500 trials . * * use lga and its competitor to solve the optimization problem @xmath184}\ ] ] with precision @xmath185 for @xmath26-argument of the global minimum on @xmath140.$ ] record the processing time for lga and its competitor . * after repeating 500 times @xmath186 calculate @xmath187 and @xmath188 the average processing time for lga and its competitor respectively . * update the file with the experimental records by adding a new line @xmath189 where @xmath61 is the degree of the polynomial . the final result is presented in the form of two curves ( average time spent versus polynomial degree ) , one for lga and the other for its competitor . total time spent @xmath190 includes all necessary supplementary steps that are needed in order to successfully implement the tested algorithm . for example , zdm total time covers calculation of critical points with laguerre s method and the subsequent search for the minimum among critical values . psm time includes calculation of the lipschitz constant or its counterparts . bfs attacks @xmath191}\ ] ] by transforming it into @xmath192 and then taking the smallest value in @xmath193 and its respective @xmath26-argument as an estimate for the solution of the optimization problem . lga outperforms bfs for polynomials with roots uniformly distributed on @xmath181\times [ -1,\;1]$ ] where @xmath194 lga considerably speeds up as the value of the parameter @xmath86 decreases ( fig . [ fig:2 ] , fig . [ fig:3 ] ) . epsf epsf if @xmath195 then lga works exactly so well as bfs ( fig . [ fig:4 ] , fig . [ fig:5 ] ) . a generic application corresponds to the situation with @xmath196 therefore employing lga instead of bfs will improve the performance of your application . epsf epsf zdm finds the global minimum @xmath191}\ ] ] by calculating zeroes @xmath197 of the derivative @xmath198\ ] ] and then finding the smallest value in @xmath199 if it is @xmath200 then zdm returns @xmath201 as an estimate for the solution of the optimization problem . in all numerical experiments reported in this paper the polynomial roots for zdm were calculated with laguerre s method @xcite , @xcite . the zdm processing time includes the invocation of laguerre s method . lga notably faster than zdm ( fig . [ fig:6 ] ) . epsf piyavskii s type algorithms tackle the optimization problem @xmath202}\ ] ] by constructing at each iteration either a piecewise linear @xmath203 is lipschitz @xcite , @xcite , @xcite , @xcite ) or a piecewise quadratic @xmath204 is lipschitz @xcite , @xcite , @xcite , @xcite ) auxiliary function @xmath205 such that @xmath206\ ] ] then the original optimization problem is replaced with @xmath207}\ ] ] based on its solution the algorithm either terminates or proceeds to the next step with the new refined auxiliary function @xmath208 lga is compared against psm with tuning of the local lipschitz constants ( referred as lt ) and its enhancement lt_li presented in @xcite . in view of the numerical simulations from @xcite lt and lt_li appear to be the fastest among pyavskii s type algorithms ( discussed in @xcite ) with piecewise linear auxiliary functions . lga is faster than lt li ( fig . [ fig:7 ] ) . epsf modifications of psm with smooth piecewise quadratic auxiliary functions are discussed in @xcite . [ fig:8 ] presents the comparison results between lga and psm enriched by local tuning of the lipschitz constant for @xmath209 ( referred as dlt ) . lga outperforms dlt . epsf dlt with some local improvement technique @xcite is addressed as dlt_li . its comparison with lga is presented by fig.[fig:9 ] . lga is faster than dlt_li . epsf the paper @xcite introduces the modification of psm based on piecewise quadratic auxiliary functions that are not necessary smooth . the algorithm from @xcite is referred in fig.[fig:10 ] as eek . lga is faster than eek . the author is grateful to anonymous referees for comments and suggestions that helped to focus and improve the original manuscript . in particular , the comparison between lga and piyavskii - shubert method was added upon a referee remark . the author is thankful to blaiklen marx for the help with preparing the manuscript for publication and testing the java implementations of algorithms in i-oblako.org framework . the program is looking for the global minimum of a polynomial on the interval @xmath1.$ ] the polynomial is represented as an array while(polynom[polynom.length - 1 ] = = 0 & & polynom.length > 1 ) { pl= new double[polynom.length - 1 ] ; for(int i = 0 ; i < polynom.length - 1;i++ ) pl[i]=polynom[i ] ; polynom = pl ; } if(polynom.length = = 1 ) return a ; if(polynom.length = = 2 ) { if(polynom[1]>= 0 ) return a ; else return b ; } pl = horner(polynom , rt_prev ) ; rt = getmin(pl , rt_prev , b , step ) ; if(rt - rt_prev<= step ) return rt_prev ; else { if(rt_prev = = a ) njumps++ ; else njumps = njumps + 2 ; } public double gradientdescent(double [ ] polynom , double a , double b , double step ) { double rt = a ; while(evalderivative(polynom , rt)<0 & & rt < b ) rt = rt+step ; return rt ; } .... d.a . adams , `` a stopping criterion for polynomial root finding . '' acm,10 , 655 - 658 , ( 1967 ) . l. armijo , minimization of functions having lipschitz continuous first partial derivatives , pacific j. math . , 16 , 1 - 3 ( 1966 ) j. m. calvin , an adaptive univariate global optimization algorithm and its convergence rate under the wiener measure , informatica , 22 , 471488 ( 2011 ) . j. m. calvin and a. zilinskas , one - dimensional global optimization for observations with noise , comput . appl . , 50 , 157169 ( 2005 ) . r. ellaia , m. z. es - sadek , h. kasbioui , modified piyavskii s global one - dimensional optimization of a differentiable function , applied mathematics , 3 , 1306 - 1320 ( 2012 ) n.j . higham , accuracy and stability of numerical algorithms , siam , philadelphia ( 2002 ) . k. hamacher , on stochastic global optimization of one - dimensional functions , phys . a , 354 , 547557 ( 2005 ) . o. gler , foundations of optimizations , springer , new york ( 2010 ) d. e. johnson , introduction to filter theory , prentice hall , new jersey , ( 1976 ) l. kantorovich and g. akilov , functional analysis in normed spaces , fizmatgiz , moscow ( 1959 ) , translated by d. brown and a. robertson , pergamon press , oxford ( 1964 ) d. kalra and a. h. barr , guaranteed ray intersections with implicit surface , comput . , 23 , 297306 ( 1989 ) d.e . knuth , art of computer programming , vol . 2 : seminumerical algorithms , 3rd ed . , ma : addison - wesley , ( 1998 ) d.e . kvasov , y.d . sergeyev , a univariate global search working with a set of lipschitz constants for the first derivative , optim . , 303 - 318 ( 2009 ) h. y .- f . lam , analog and digital filters - design and realization , prentice hall , new jersey , ( 1979 ) yu . nesterov , introductory lectures on convex optimization , applied optimization , 87 , kluwer academic publishers , boston ( 2004 ) s. piyavskii , an algorithm for finding the absolute minimum of a function , theory of optimal solutions , ik akad . nauk ussr , kiev , 2 , 13 - 24 , ( 1967 ) s. piyavskii , an algorithm for finding the absolute extremum of a function , ussr comput . , 12 , 57 - 67 , ( 1972 ) b. shubert , a sequential method seeking the global maximum of a function , siam j. numer . , 9 , 379 - 388 ( 1972 ) d. lera and y. d. sergeyev , acceleration of univariate global optimization algorithms working with lipschitz functions and lipschitz first derivatives , siam j. optim . , 23 , 1 , 508 - 529 ( 2013 ) y. d. sergeyev , global one - dimensional optimization using smooth auxiliary functions , math . programming , 81 , 127 - 146 ( 1998 )
the paper proposes a new algorithm for solving global univariate optimization problems . the algorithm does not require convexity of the target function . for a broad variety of target functions after performing ( if necessary ) several evolutionary leaps the algorithm naturally becomes the standard descent ( or ascent ) procedure near the global extremum . moreover , it leads us to an efficient numerical method for calculating the global extrema of univariate real analytic functions .
the recent observation by lhcb @xcite of prominent exotic structures @xmath2 and @xmath3 in the @xmath7 spectrum of @xmath10 has rekindled hopes that the long - sought pentaquark states have finally been observed . the reported properties are @xmath11 mev , @xmath12 mev ( at 9 standard deviations ) and @xmath13 mev , @xmath14 mev ( at 12 standard deviations ) , respectively , while the preferred @xmath15 assignments are correlated , with the most likely combinations in decreasing order being @xmath16 , @xmath17 , and @xmath18 . should at least one of the states be confirmed by another experiment , it will join the famed tetraquarks whose best - studied member [ the @xmath19 @xmath20 was discovered over a decade ago @xcite as a second class of exotic hadrons . since the valence structure of @xmath7 is @xmath21 , the minimal quark content of such states is that of a pentaquark . we note several interesting phenomenological facts . first , the two charged _ five_-quark hidden - charm states @xmath2 and @xmath3 observed at lhcb are both lighter than the _ four_-quark state @xmath22 [ formerly @xmath23 observed by the same group @xcite . since pentaquarks carry baryon number and therefore must have a baryon in their decay products , while tetraquarks are bosons and therefore can decay entirely to ( generically lighter ) mesons , less phase space is typically available to @xmath6 decays . in particular , @xmath22 decays dominantly to @xmath24 , even though plenty of phase space is available for the @xmath25 mode , while neither of the @xmath6 states is kinematically allowed to decay to @xmath26 . second , both of the @xmath6 states lie well above the thresholds for decay into @xmath27 , which also has the valence - quark structure @xmath21 . since these two - body decays are not forbidden by any obvious quantum number , the relatively small @xmath6 widths suggest interesting internal structure sufficient to suppress this immediate rearrangement , as well as that into @xmath28 . the @xmath3 lies only about 10 mev below the @xmath29 threshold , which can be used to argue for a molecular interpretation @xcite . alternate molecular assignments for the @xmath6 states are presented in refs . @xcite , and the origin of the @xmath3 as a threshold rescattering effect via @xmath30 is discussed in ref . @xcite , and through additional channels in ref . @xcite . third , all of the preferred fits from lhcb demand that the two @xmath6 states carry opposite parities . a system of four quarks and one antiquark in a relative @xmath31 wave has negative parity , while positive parity requires the introduction of at least one unit of relative orbital angular momentum ( @xmath32 wave or higher ) . and yet , the two @xmath6 states are separated by only @xmath33 mev . one may argue that the figure of merit , as would be the case for regge trajectories , should be @xmath34 , which is a much more natural hadronic scale ; however , attempting to discern a trajectory when only two points are available seems absurdly premature . still , if the @xmath35 state @xmath36 seen by belle with a mass @xmath37 mev @xcite is confirmed , then small mass splittings between hidden - charm exotics of opposite parities [ _ i.e. _ , only 20 mev from the @xmath20and hence relative orbital excitations will not appear unusual . in this short paper , we propose a structure for the @xmath6 states based upon a mechanism recently proposed for understanding the tetraquark states @xcite , wherein a diquark @xmath38 and an antidiquark @xmath39 form in the attractive color-@xmath40 and color-@xmath41 representations , respectively . this configuration is prevented from instantly reorganizing into two @xmath42 pairs because the diquarks are forced to separate rapidly due to the large energy release in the production mechanism ( _ e.g. _ , a @xmath43-meson decay via @xmath44 ) . however , since the diquarks are colored , they are confined and hence can not separate indefinitely ; kinetic energy is converted into the potential energy of a color flux tube connecting them . hadronization finally occurs through the overlap of the long - distance tails of hadronic wave functions that stretch from the quarks in @xmath38 to the antiquarks in @xmath45 . this picture was used to explain , for example , the preference for @xmath46 over @xmath47 , simply due to the final spatial separation of the @xmath48 in @xmath38 and the @xmath49 in @xmath39 allowing a much greater wave function overlap with the larger @xmath50 than with the more compact @xmath51 . we argued implicitly in @xcite , and explicitly in a subsequent paper @xcite , that the diquarks formed from one heavy and one light quark are somewhat smaller than diquarks formed from two light quarks , and that such heavy - light @xmath52 diquarks are not expected to be much larger than the meson @xmath53 formed from the same flavors . in yet another subsequent paper @xcite , we proposed that the preferential coupling of @xmath54 ( and its conjugate ) does not end with diquarks , but can continue sequentially to more complex structures like pentaquarks and even octoquarks . these ideas will be used to develop the pentaquark picture described in detail below . as in ref . @xcite , we use the term `` picture '' because many distinct models could be constructed that support this dynamics , and want to emphasize that this discussion is not limited to any particular choice of potential or wave functions , for example . the concept that pentaquarks might be formed from two compact colored constituents rather than molecules of mesons was first expressed in ref . @xcite , which sought to describe the @xmath55 pentaquark candidate @xmath56 as a molecule of a @xmath57 diquark in a color-@xmath58 and a @xmath59 `` triquark '' ( a term coined in that paper ) formed from a @xmath57 diquark in a color-@xmath60 coupled to the @xmath61 quark into an overall color-@xmath41 . the separation of the two components of the molecule is stabilized by the centrifugal barrier introduced by relative orbital angular momentum @xmath62 , which as discussed above was necessary should the @xmath63 have been found to carry positive parity . in comparison , the famous @xmath63 pentaquark model of ref . @xcite proposed a structure of two light @xmath57 diquarks and one exceptional ( @xmath61 ) quark . both refs . @xcite and @xcite , make use of diquarks consisting of light quarks in the `` good '' ( spin-0 ) combinations , so named because they are believed to be more tightly bound than the `` bad '' ( spin-1 ) combination through hyperfine couplings ( although in ref . @xcite the diquark inside the triquark has spin 1 ) . in the case of heavy - light diquarks , the hyperfine couplings are proportional to @xmath64 , and therefore the mass difference between `` good '' and `` bad '' is greatly reduced . alternate compositions through colored components have been very recently discussed in refs . @xcite . this paper is organized as follows : in sec . [ sec : diquark ] we discuss the diquark picture used to describe the new exotic states . section [ sec : pentapic ] presents the diquark picture as relevant to the production of the @xmath6 states . in sec . [ sec : phenom ] we present the basic phenomenology for the @xmath6 states provided by this picture , and in sec . [ sec : concl ] we summarize . the color algebra of qcd provides more than one way to obtain attractive channels between quarks . the combination @xmath65 , which explains the strong binding of @xmath66 pairs in conventional mesons , is of course extremely well known . however , one other binary combination is strongly binding , the channel @xmath54 . to quantify the effect , note that the coupling of two colored objects in the irreducible representations @xmath67 and @xmath68 is computed by the same techniques as one computes the spin coupling between objects carrying spins @xmath69 and @xmath70 combining to total spin @xmath71 , via the trick @xmath72 \ , .\ ] ] one generalizes to an arbitrary lie algebra by computing the combination of quadratic casimirs @xmath73 considering all binary combinations @xmath74 and @xmath75 , one computes the relative strengths @xmath76 the diquark attractive @xmath58 channel @xmath38 is therefore fully half as strong as that of the singlet @xmath77 channel . in a multiquark system in which two quarks or two antiquarks happen to lie in closer proximity than to one of their antiparticles , the diquark @xmath38 ( antidiquark @xmath39 ) attraction is naturally expected to dominate . unless stronger color forces intervene ( _ e.g. _ , the production of nearby @xmath78 s , which would create available color - singlet combinations ) , the @xmath38 and @xmath39 combinations can be expected to form quasi - bound , but colored and therefore confined , states . in ref . @xcite , the diquarks containing a charm and a light quark were crudely estimated to have a comparable size to a @xmath79 meson , roughly @xmath80 fm . the subsequent paper ref . @xcite argued that diquarks formed with a heavy quark should be somewhat smaller than those formed from two light quarks : a heavier quark is more localized in space , while each lighter quark has a more diffuse wave function . a key phenomenological question in identifying whether diquarks have formed is whether or not any antiquarks appear within this radius . this hypothesis was used in ref . @xcite to suggest a means by which multiquark exotics could be produced , particularly at threshold , where the limited phase space allows the soft heavy quark pairs such as @xmath81 to coalesce with light valence quarks moving at similar rapidities . such multiquark states can be formed through a sequence of two - body bound - state clusters of color-@xmath58 diquark and color-@xmath41 antidiquark states . in the absence of easy opportunities for the formation of color singlets , sequential diquark formation provides the strongest channels for binding . for example , anti - de sitter / qcd models on the light front @xcite have a universal confining potential that confirms the importance of diquarks in hadron spectroscopy . while the examples given in ref . @xcite describe literal clusters of diquarks such as a charmed , charge @xmath82 , baryon - number @xmath83 state @xmath84_{\bar 3c } [ cu]_{\bar 3c } [ uu]_{\bar 3c}$ ] , another route of sequential color - triplet ( antitriplet ) formation is available , which is the hypothesis of this paper : a pre - existing diquark @xmath85 that subsequently encounters an antiquark @xmath86 forms a bound _ antitriquark _ @xmath87 via the attractive color coupling @xmath88 . this mechanism , as discussed in the next section , provides a completely analogous production channel for pentaquark states to that described for tetraquark states in ref . @xcite . to say that two quarks or antiquarks encountering one another combine only into the most attractive channel is of course a great simplification . first , the color coupling factors apply without reservations only when fundamental qcd interactions dominate the interaction . longer - distance effects dress the interactions and can obfuscate this simple result . in reality , one expects a type of thermodynamic ensemble of states in various color combinations , in which the levels at the lowest energies are driven by diquark binding . the existence of such an ensemble assumes that the formation of overall color singlets is precluded due to the presence of a potential energy barrier , such as from a large spatial separation between the quarks needed to form the singlets . in that case , the eventual hadronization can be considered as a tunneling process . second , the pauli exclusion principle must be taken into account if the purported diquarks contain quarks of identical flavor , since then the flavor wave function is automatically symmetric , and therefore the color - spin wave function must be antisymmetric . in the specific example discussed below , this constraint is not an issue , but it must be kept in mind for other cases . we propose that the states @xmath6 observed at lhcb are pentaquarks consisting of a confined but rapidly separating pair of a color-@xmath40 diquark @xmath89 and a color-*3 * antitriquark @xmath90 , in which the @xmath57 subsystem of @xmath91 is a color-@xmath58 diquark @xmath85 , as depicted in fig . [ fig : pc ] . attracted by long - range color forces ( indicated by gluon lines ) via a color flux tube . here , the mechanism is illustrated for @xmath92 , where the black square indicates the @xmath93-quark weak decay.,width=345 ] this picture is completely analogous to ( the charge conjugate of ) that for the @xmath94 presented in ref . @xcite , except that the diquark @xmath95 in fig . [ fig : pc ] is replaced by the single quark @xmath96 . the parent hadron for the @xmath6 is the @xmath97 baryon , while the parent hadron for the @xmath98 is the @xmath99 . in either case , the composite state is not a molecule in the traditional sense of the word , because it lasts only as long as the @xmath38-@xmath100 pair ( @xmath38-@xmath39 in ref . @xcite ) possess positive kinetic energy to continue separating . the colored @xmath38 and @xmath100 constituents create a color flux tube between them , losing their energy to the color field and slowing down . for this picture to be physically meaningful , the @xmath38-@xmath100 pair must be sufficiently compact that their wave function overlap becomes insignificant . as in the standard wentzel - kramers - brillouin approximation , the probability of a transition in this case , hadronization increases as the components approach the classical turning point . let us emphasize the differences between this and other pictures for the pentaquark . first , it is clearly different from the hadronic molecule picture , in which the constituent baryon and meson are color singlets treated as forming a static molecule and are held together by weak color van der waals forces . it is also different from the diquark - triquark model of ref . @xcite , not only because the diquarks are always assumed here to form color triplets , but also because the diquark - triquark state in that case is again a static molecule , stabilized by a centrifugal barrier ; in our picture , such states are expected to last only as long as the components continue to separate , and also would exist in @xmath31 waves as well as higher partial waves . the recent work of ref . @xcite describes the @xmath6 states as being formed of diquarks , via the composition @xmath101 ; while the color structure is the same as ours , we emphasize the importance of the @xmath49 belonging to a compact component of the overall state . the new ingredient for the @xmath6 as compared to the @xmath94 is the use of the intrinsic diquark @xmath95 originating in the @xmath97 . note from fig . [ fig : pc ] that @xmath85 acts as a spectator in the @xmath6 production process . @xmath102 baryons have always had a special place in the history of diquark models , because these baryons by definition are isosinglets , and since the heavy quarks @xmath103 are also isosinglets , the remaining quark pair @xmath57 also forms an isosinglet , with a wave function that is antisymmetric under flavor exchange . since @xmath104 is a color-*3 * , @xmath57 is a color @xmath58 , again an antisymmetric combination . the pauli principle therefore demands an antisymmetric spin wave function for @xmath57 . since both the ground - state @xmath102 baryons and the heavy quarks @xmath104 have @xmath105 , the @xmath57 is therefore expected to live in the antisymmetric spin-0 combination . the @xmath57 pair in @xmath102 baryons is frequently termed a diquark , and indeed it has exactly the color structure we want ; it differs from the ones we have previously discussed only by consisting solely of light quarks , and is in the `` good '' diquark combination only . as mentioned above , one may expect the @xmath57 diquark to be slightly larger than the heavy - light diquarks , but even so , its binding to the heavy quark @xmath93 in @xmath97 and @xmath48 in @xmath100 restricts the full spatial extent of the wave function . for example , using heavy - quark symmetry and a variational approach , ref . @xcite calculate root - mean square matter radii for @xmath97 and @xmath106 to be no more than 0.22 fm and 0.31 fm , respectively . treating the @xmath100 antitriquark as a `` would - be '' @xmath106 baryon ( _ i.e. _ , differing by @xmath107 but otherwise bound by essentially the same nonperturbative physics ) , one expects @xmath100 to be not much larger than @xmath106 . in principle , the @xmath108 quark created from a gluon can mix with the one in the @xmath97 diquark . however , inasmuch as this initial @xmath57 diquark is expected to be fairly tightly bound , one expects it to propagate as an undisturbed spectator quasiparticle through the process ; otherwise , the most likely outcome would be a dissociation of the diquark , leading to a different intermediate state than described in fig . [ fig : pc ] . in the case of the observed decay @xmath109 , the decay rate is suppressed due to the final separation of the @xmath48 quark ( in @xmath38 ) and @xmath49 quark ( in @xmath100 ) compared to the typical size of the @xmath51 wave function , and to a lesser extent due to the separation of the @xmath57 diquark @xmath85 in @xmath91 from the @xmath108 quark in @xmath38 compared to the typical size @xmath110 fm of the proton wave function . the first interesting point one notes from fig . [ fig : pc ] is the peripheral role played in the process by the @xmath111 pair created by gluodynamics . certainly , creation instead of a @xmath112 pair would give a nearly identical scenario . one therefore predicts isodoublet partners @xmath9 to be produced via @xmath113 at masses just a few mev higher ( from @xmath114 ) than those for the @xmath6 states . that the @xmath115 states should form isospin doublets is of course guaranteed by @xmath116 being a strong decay and hence conserving isospin , while @xmath117 and @xmath118 ; nevertheless , it is useful to see the charge symmetry process explicitly in the context of a particular decay mechanism . we now obtain a crude estimate of the separation of the diquark and antitriquark using the same technique as in ref . @xcite : since the two components transform as a color-@xmath119 pair , one may describe them using the well - known linear - plus - coulomb `` cornell '' nonrelativistic potential @xcite . in the most thorough recent analysis @xcite , the central part of the potential for @xmath81 systems is given by @xmath120 with @xmath121 , @xmath122 gev@xmath123 , @xmath124 gev , and @xmath125 gev . the @xmath126 color factor is the same one as in eq . ( [ eq : colors ] ) , and @xmath127 now refer to the components containing these quarks , in our case @xmath38 and @xmath91 , respectively . the calculation of ref . @xcite further exploited the fact that at least one of the components @xmath38 and @xmath39 is in a state of zero spin for each state of interest [ @xmath128 and @xmath22 ] , so that @xmath130 , and that both are expected to have all quarks in a relative @xmath31 wave , so that noncentral contributions to @xmath131 are not needed . in the case of the @xmath6 states , we have seen that at least one of them must have an orbital excitation , so this calculation strictly applies only to the @xmath31-wave state . furthermore , the antitriquark @xmath100 necessarily carries half - integer spin ; nevertheless , the @xmath132 term has little effect on @xmath131 except at the smallest values of @xmath133 , so we neglect it here . lastly , we use the qcd sum - rule based estimate @xcite @xmath134 mev ( note its nearness to the @xmath135 mass 1.865 gev ) and the antitriquark mass estimate @xmath136 gev . using these assumptions , the diquark - antitriquark separations @xmath137 obtained from eq . ( [ eq : cornell ] ) are @xmath138 these distances are not especially large for light hadronic systems . however , inasmuch as diquarks , and especially triquarks , containing heavy quarks may be rather smaller as discussed above , these components may be considered as well separated for the purpose of computing quantum - mechanical wave function overlaps . this separation , particularly of the @xmath48 and @xmath49 quark , explains the suppressed decay rate to @xmath51 , since the potential eq . ( [ eq : cornell ] ) gives @xmath139 fm . the similarly small size for the @xmath106 also predicts slow transitions to @xmath140 , and hence overall widths that are suppressed compared to naive expectations . finally , let us consider the quantum numbers of the allowed states in this picture . orbital excitations can occur not only along the flux tube , but within the diquark and antitriquark as well . nevertheless , let us for simplicity ignore the latter . inasmuch as the diquark @xmath85 in @xmath100 inherited from the @xmath97 has spin zero , the set of allowed quantum numbers is even simpler , since then the spin of @xmath100 is @xmath141 . for @xmath31 waves , one has the @xmath15 possibilities @xmath142 and for @xmath32 waves , one has @xmath143 where the three numbers on the left - hand side are the spin of the @xmath144 antitriquark , the orbital excitation , and the spin of the @xmath145 diquark . in this simplified picture , we find only one state with @xmath146 , namely , the @xmath32-wave @xmath147 . it is natural to identify this state with the higher - mass @xmath3 , since it has a narrower width that can be explained by the near - threshold phase - space suppression of @xmath32 waves . then the broader @xmath2 must have @xmath148 and lie in an @xmath31 wave . note that both of these states have the @xmath145 diquark in a spin-1 configuration . it is an interesting phenomenological fact of the tetraquark sector that no @xmath149 state has yet been confirmed ; in the context of the diquark - antidiquark picture , the simplest such states would have both diquarks in @xmath150 combinations , with @xmath151 in the color flux tube as well . it is plausible that such states are much broader due to the absence of any angular momentum barriers impeding rapid decays . likewise , the lower - spin states in eqs . ( [ eq : swaves ] ) and ( [ eq : pwaves ] ) might be more difficult to discern experimentally . in any case , the discovery of two new states and the possibility of numerous others left to find will certainly spur on further experimental examination . we have seen that the recently observed charmoniumlike pentaquark candidates @xmath6 may have a common dynamical origin with the charmoniumlike tetraquark states . both are proposed to occur as systems of rapidly separating color-*3 * and -@xmath58 component pairs , and in particular in the @xmath6 pentaquarks through the sequential preferential formation of color - triplet combinations @xmath152^{\vphantom{(}}_{\bf 3 } ( cu)_{\bar{\bf 3}}$ ] . the diquark @xmath145 and antitriquark @xmath144 achieve a substantial separation before hadronization must occur , providing a qualitative explanation for the suppression of the measured widths compared to available phase space . the @xmath6 states in this picture form isospin doublets with neutral , as - yet undiscovered partners . states with the observed @xmath15 quantum numbers can easily be accommodated in this scheme , and suggest the potential for discovery of numerous additional related states in the future . i am deeply indebted to s. brodsky for his insights and encouragement , and also thank m. karliner for interesting discussions . this work was supported by the national science foundation under grant nos.phy-1068286 and phy-1403891 . 99 r. aaij _ et al . _ [ lhcb collaboration ] , arxiv:1507.03414 [ hep - ex ] . choi _ et al . _ [ belle collaboration ] , phys . lett . * 91 * , 262001 ( 2003 ) [ hep - ex/0309032 ] . r. aaij _ et al . _ [ lhcb collaboration ] , phys . rev . lett . * 112 * , 222002 ( 2014 ) [ arxiv:1404.1903 [ hep - ex ] ] . m. karliner and j.l . rosner , arxiv:1506.06386 [ hep - ph ] . l. roca , j. nieves , and e. oset , arxiv:1507.04249 [ hep - ph ] . yang , z.f . sun , j. he , x. liu , and s.l . zhu , chin . c * 36 * , 6 ( 2012 ) [ arxiv:1105.2901 [ hep - ph ] ] . r. chen , x. liu , x.q . li , and s.l . zhu , arxiv:1507.03704 [ hep - ph ] . chen , w. chen , x. liu , t.g . steele , and s.l . zhu , arxiv:1507.03717 [ hep - ph ] . j. he , arxiv:1507.05200 [ hep - ph ] guo , u.g . meiner , w. wang , and z. yang , arxiv:1507.04950 [ hep - ph ] . liu , q. wang , and q. zhao , arxiv:1507.05359 [ hep - ph ] . liu _ et al . _ [ belle collaboration ] , phys . lett . * 110 * , 252002 ( 2013 ) [ arxiv:1304.0121 [ hep - ex ] ] . brodsky , d.s . hwang , and r.f . lebed , phys . lett . * 113 * , 112001 ( 2014 ) [ arxiv:1406.7281 [ hep - ph ] ] . s.h . blitz and r.f . lebed , phys . d * 91 * , 094025 ( 2015 ) [ arxiv:1503.04802 [ hep - ph ] ] . s.j . brodsky and r.f . lebed , phys . d * 91 * , 114025 ( 2015 ) [ arxiv:1505.00803 [ hep - ph ] ] . m. karliner and h.j . lipkin , phys . b * 575 * , 249 ( 2003 ) [ hep - ph/0402260 ] . jaffe and f. wilczek , phys . lett . * 91 * , 232003 ( 2003 ) [ hep - ph/0307341 ] . a. mironov and a. morozov , arxiv:1507.04694 [ hep - ph ] . l. maiani , a. d. polosa , and v. riquer , arxiv:1507.04980 [ hep - ph ] . brodsky , g.f . de teramond , h.g . dosch , and j. erlich , phys . rept . * 584 * , 1 ( 2015 ) [ arxiv:1407.8131 [ hep - ph ] ] . c. albertus , j.e . amaro , e. hernandez , and j. nieves , nucl . a * 740 * , 333 ( 2004 ) [ nucl - th/0311100 ] . e. eichten , k. gottfried , t. kinoshita , k.d . lane , and t .- m . yan , phys . d * 17 * , 3090 ( 1978 ) [ erratum - ibid . d * 21 * , 313 ( 1980 ) ] . t. barnes , s. godfrey , and e.s . swanson , phys . d * 72 * , 054026 ( 2005 ) [ hep - ph/0505002 ] . kleiv , t.g . steele , a. zhang , and i. blokland , phys . d * 87 * , 125018 ( 2013 ) [ arxiv:1304.7816 [ hep - ph ] ] .
starting with the dynamical picture of the exotic @xmath0-containing states @xmath1 as the confinement - induced hadronization of a rapidly separating pair of a compact diquark and antidiquark , we describe the pentaquark candidates @xmath2 and @xmath3 in terms of a confined but rapidly separating color - antitriplet diquark @xmath4 and color - triplet `` triquark '' @xmath5 . this separation explains the relatively small @xmath6 widths , despite these 5-quark systems lying far above both the @xmath7 and @xmath8 thresholds . the @xmath6 states are predicted to form isospin doublets with neutral partners @xmath9 .
suppose that @xmath0 is a compact oriented @xmath1manifold without boundary . for any lie algebra , @xmath2 , and integral level , @xmath3 , there is defined an invariant , @xmath4 , of embeddings of links @xmath5 in @xmath0 , known as the witten - reshetikhin - turaev invariant ( see @xcite , @xcite ) . it is known that for links in @xmath6 , @xmath7 is a polynomial in @xmath8 , namely the generalized jones polynomial of the link @xmath5 . in this paper we consider only the case @xmath9 . now assume that @xmath0 is a rational homology sphere , with @xmath10 . in the normalization for which the invariant for @xmath6 is @xmath11 , denote the invariant for the pair @xmath12 , as an algebraic function of @xmath13 at @xmath14 roots of unity , by @xmath15 . for a rational homology sphere @xmath0 and odd prime @xmath16 , @xcite showed that @xmath17 $ ] , so that for some @xmath18 , one has @xmath19 although the @xmath20 are not uniquely determined , it is known from @xcite that there exist rational numbers @xmath21 such that , @xmath22 as elements of @xmath23 for all sufficiently large primes @xmath16 . for integer homology spheres , @xmath24 and @xmath25 where @xmath26 denotes the casson walker invariant of @xmath0 in casson s normalization . as a result , one may define a formal power series @xmath27 with rational coefficients , which is an invariant of rational homology @xmath1spheres , @xmath0 , known as the _ ohtsuki series _ of @xmath0 . this is the asymptotic expansion of the trivial connection contribution to @xmath15 ( see @xcite ) . work of habiro ( see @xcite , @xcite ) showed that @xmath28 can always be written in the form @xmath29 where @xmath30 $ ] . this expression has the advantage that at a root of unity it reduces to a finite sum ( polynomial in @xmath31 , while it also can be considered as an element of @xmath32 $ ] . in this paper we review the technique used by le @xcite and habiro @xcite , using the @xmath33-matrix presentation of link and manifold invariants from @xcite , @xcite , to compute the colored jones function for a knot and hence the ohtsuki series for integer homology spheres obtained by surgery around a knot . we carry out the quantum group calculation in general , obtaining the colored jones polynomial of a knot as a state sum directly in terms of @xmath13-numbers as in @xcite ( though here it is written for arbitrary knot diagrams , not just for braid closures ) . for arbitrary integer homology spheres the result would be similar , but the knot case is simpler , since then the gauss diagram involves only one circle . an outline of the present paper is as follows . in 2 , basic notation is defined relating to @xmath13-numbers along with a summary of facts needed concerning the quantum group @xmath34 . in 3 , we follow @xcite and le to give a description of the @xmath35 colored jones polynomial of a knot , first in terms of @xmath33-matrices , then using the universal invariant as an algebra element and finally evaluating to get the state sum formula ( theorem 1 ) for the colored jones function based on a gauss diagram description of the knot . in 4 , following le and habiro , the transformation from the colored jones function of a knot to the ohtsuki series of the 3-manifolds obtained by surgery around the knot , is discussed and combining with theorem 1 gives habiro s cyclotomic form for the ohtsuki series . this formulation is used to give bounds on the growth of coefficients in the ohtsuki series , which support the slope conjecture of @xcite . finally in 5 , the application of the formulae to the trefoil and figure-8 knots are demonstrated . let @xmath37 be a formal parameter . define @xmath13-numbers , @xmath13-factorials and @xmath13-binomial coefficients according to @xmath38={v^n - v^{-n}\over{v - v\inv}}\>,\qquad[n]!=\prod_{i=1}^n{[i]}\ > , \qquad\binomial{n}{m}={[n]!\over[m]![n - m]!}\>.\ ] ] the dependence on @xmath13 will be omitted from the notation . we will also define asymmetric @xmath13-numbers , @xmath13-factorials and @xmath13-binomial coefficients according to @xmath39 there is an inductive relation as for ordinary binomial coefficients @xmath40 from which it follows that the asymmetric @xmath13-numbers , @xmath13-factorials and @xmath13-binomial coefficients all lie in @xmath41 $ ] . also , even when @xmath42 is not an integer , say an element of an algebra , we can make sense of quantum binomial coefficients of the form @xmath43\over[i]}\ > , \qquad\binomials{x}{m}=\prod_{i=1}^m{\{x - i+1\}\over\{i\}}\>,\ ] ] where @xmath44 . the two sets of @xmath13-numbers are related by @xmath45\>,\qquad\{n\}!=v^{{1\over2}n(n-1)}[n]!\ > , \qquad\binomials{n}{m}=v^{m(n - m)}\binomial{n}{m}\>.\ ] ] we will also denote by @xmath46 , the asymmetric @xmath13-number obtained by replacing @xmath13 by @xmath47 , and by @xmath48 the corresponding @xmath13-factorial . thus @xmath49 and @xmath50 . the quantum group @xmath52 is defined by generators @xmath53 , @xmath54 and @xmath55 with relations @xmath56=[h]\>,\qquad[h , x]=2x\>,\qquad[h , y]=-2y\>.\ ] ] set @xmath57 . the comultiplication @xmath58 is given by @xmath59 with antipode @xmath60 being an antihomomorphism acting on the generators by @xmath61 additionally @xmath62 is a quasi - triangular hopf algebra with universal @xmath33-matrix , @xmath63 satisfying the yang - baxter equation @xmath64 . a formula for @xmath33 ( see @xcite ) is @xmath65 write this as @xmath66 ( thus the suffix @xmath67 refers to a pair @xmath68 of non - negative integers ) . as in any quasi - triangular hopf algebra , @xmath69 , @xmath70 we will write these two formulae together for @xmath71 as @xmath72 the square of the antipode is then given by @xmath73 where @xmath74 and @xmath75 is central . furthermore , @xmath62 is a ribbon hopf algebra , that is there is a central ( ribbon ) element @xmath76 such that @xmath77 in our case , @xmath78 . this special ( charmed ) element satisfies the property @xmath79 , while @xmath80 lies in the center of @xmath62 . for each @xmath81 , there is a @xmath82-dimensional module @xmath83 with basis @xmath84 . the quantum group action is described by the lowest weight vector @xmath85 for which ( we omit the superscripts @xmath86 ) @xmath87 since @xmath88=[i][h - i+1]x^{i-1}$ ] , the whole action is given by @xmath89[\mu - i]v_{i-1}\>.\ ] ] since @xmath90 , it acts as multiplication by a scalar in any irreducible representation . in particular , in @xmath83 , @xmath91 acts by scaling by @xmath92 as can be verified by direct evaluation on @xmath93 , where @xmath53 and @xmath16 act as @xmath94 and @xmath95 respectively . suppose @xmath5 is an oriented framed link in @xmath6 . the _ generalized colored jones polynomial _ @xcite of @xmath5 is defined when each component @xmath96 is colored by a representation @xmath97 of @xmath62 and will be denoted by @xmath98 . we will assume for simplicity that @xmath97 are irreducible . indeed , according to @xcite , there is a functor , @xmath99 , from the category of ( colored ) ribbon tangles to the category of vector spaces , under which links ( closed tangles ) map to scalars , namely @xmath100 . a slice of a colored ( oriented ) ribbon tangle ( object in the category ) is an ordered list of colors @xmath101 with orientations @xmath102 and is mapped under the functor @xmath99 to @xmath103 , where @xmath104 denotes the dual representation @xmath105 . this functor is defined by its images on the generators shown in figure 1 . in our conventions , downward oriented strands are counted positively in objects , while tangle morphisms are composed considering them from bottom to top . + the two orientations of crossing , in which the strands are labeled with representations @xmath106 and @xmath107 as shown , have as images the maps @xmath108 given by @xmath109 and @xmath110 , respectively , where @xmath111 is the permutation of the factors . the four cup and cap sections transform to maps @xmath112 and @xmath113 respectively . here @xmath114 is the dual basis for @xmath105 to the basis @xmath115 for @xmath106 . suppose that @xmath5 has one component ( knot ) and @xmath116 is a 1-tangle presentation of @xmath5 , that is @xmath117 ( where @xmath118 is the object consisting of one downward oriented point ) such that its closure with blackboard framing is ambient isotopic to @xmath5 . then the above prescription can be used to compute @xmath119 ( when @xmath118 and @xmath116 are colored by @xmath106 ) as a map @xmath120 , for any representation @xmath106 ( the color of the one open strand ) . this map commutes with the action of @xmath62 , so that if @xmath106 is irreducible it is given by multiplication by a scalar , namely @xmath121 where @xmath122 is the unknot with framing zero . for @xmath122 we have @xmath123\>.\ ] ] by construction , @xmath124 is multiplicative under connect sum on @xmath5 ( corresponding to composition of tangles @xmath116 ) . in particular , the algebra element corresponding to a trivial 1-tangle with framing 1 is @xmath91 , so that @xmath124 changes by a factor @xmath125 under a framing change of @xmath126 in @xmath5 . the above functorial prescription for @xmath124 when @xmath5 is a knot expressed as the closure of a 1-tangle @xmath116 , may be rewritten algebraically as follows . first assume that @xmath116 is presented as a tangle diagram in generic position , in which the crossings are between downward oriented strands ( this is always possible ) . recall that @xmath128 where the sum is over @xmath129 an indexing set ( in this case , pairs of non - negative integers @xmath68 ) . place an element of @xmath130 at each crossing . in the neighborhood of each crossing decorate the crossing stands by elements of @xmath62 according to figure 2 ; @xmath131 on the overcrossing arc and @xmath132 on the undercrossing arc , where @xmath133 denotes the sign of the crssing ( @xmath134 for a slash crossing and @xmath135 for a backslash crossing ) . at local maxima / minima oriented leftwards decorate the strand by @xmath136 , @xmath137 respectively ; rightward oriented cups and caps receive no decoration . + then @xmath138 is the evaluation in the chosen representation of the algebra element read off the diagram by tracing the tangle strand according to its orientation , and writing down the decorations in the form of a product from left to right , and then summing over all labels @xmath67 at crossings . in fact , in the correct quotient of @xmath62 , the algebra element itself is an invariant ( this is the universal @xmath139 invariant ) ; however we do not need this here . in a general tangle diagram , crossings need not always occur only between downward oriented strands . however , other crossings can always be redrawn in terms of such downward oriented crossings , at the cost of introducing extra caps and cups , by internally rotating " the crossing . for example a crossing of two upward oriented strands can be obtained by internally rotating " a downward oriented crossing through @xmath140 , either clockwise or anticlockwise and consistency of the results follows from the fact that @xmath141 along with @xmath142 @xmath143 . + there are altogether eight crossing types given by the orientations on the two strands as well the sign of the crossing itself . the result is that for an arbitrarily oriented crossing , the undercrossing arc is decorated with @xmath144 while the overcrossing arc is decorated with @xmath145 _ except _ in the case of a left pointing crossing in which case it is decorated by @xmath146 . + we will now proceed to carry out the above prescription at the quantum group level for an arbitrary 1-tangle , producing a prescription for @xmath147 as a state sum of scalar quantities . the combinatorial data involved in specifying a knot @xmath5 , presented as a 1-tangle closure can be contained in its gauss diagram . that is , consider the knot as parametrised by an ( orieted ) circle . each crossing in @xmath5 corresponds to exactly two points on the circle , one for the underpass and the other for the overpass . encode this by joining the points on the circle by an oriented chord from overpass to underpass , marked by a sign to denote the sign of the crossing . the resulting diagram is known as the _ gauss diagram _ of @xmath5 ; it is an oriented signed chord diagram on an oriented circle . this is now sufficient data to encode the knot . unfortunately the above description of the algorithm for @xmath147 requires also a tangle presentation with left pointing crossings / cups / caps marked . this requires that the circle be marked with a basepoint as well as blobs @xmath148 at left pointing cups / caps with additionally left pointing crossings marked . ( note that only left pointing cups / caps / crossings in the long knot are marked ; these may differ from those in the closure ) . since @xmath149 , instead of marking a chord as left pointing " ( whose corresponding crossing was left pointing ) , one may equivalently place @xmath150 and @xmath151 blobs on the outer circle just before and after the outward pointing ( corresponding to the overcrossing arc ) end of that chord . the result is an _ enhanced gauss diagram _ in which additional marks have been placed on the outer circle , namely the basepoint and blobs @xmath148 . figure 5 shows an example for the figure 8 knot . + let @xmath152 be such an enhanced gauss diagram for a knot @xmath16 with @xmath153 crossings . in order to write formulae more easily , we give complete combinatorial data for @xmath152 . number the crossings ( chords of @xmath152 ) by @xmath154 . let @xmath155 denote the sign ( orientation ) of the crossing , positive for a slash crossing pointing up / down . as we go around the circle , we encounter @xmath156 contributions from crossings which we will index by @xmath157 ; for each such @xmath67 there is a pair @xmath158 giving the crossing number @xmath159 and the sign with which it is encountered , @xmath160 , which is positive for an overcrossing and negative for an undercrossing . finally let @xmath161 denote the positions of the blobs @xmath148 , the @xmath162 blob appearing between the @xmath163 and @xmath164 positions on the circle , with contribution @xmath165 . throughout the remainder of this paper , we will always use @xmath166 as an index for crossings and @xmath67 , @xmath3 as indices for vertices on the gauss diagram . the evaluation of @xmath119 in the representation @xmath167 may be carried out by computing the action on the lowest weight vector @xmath168 ( equivalently it is the @xmath169 matrix element ) , and then @xmath170 here @xmath171 is the @xmath169 matrix element in the evaluation on @xmath83 of the quantum group word @xmath172 obtained by tracing round the outer circle of @xmath152 starting at the basepoint , with contributions of @xmath173 ( outward arrow / overcrossing ) , @xmath174 ( inward arrow / undercrossing ) and @xmath148 ( blobs on the circle ) , written in order from left to right . it is convenient to use the notation @xmath175 and @xmath176 . using @xmath177 we push all powers of @xmath54 and @xmath55 to the end of the word , obtaining @xmath178 where @xmath179 the product being written from left to right . ( note that @xmath180 is independent of @xmath181 . ) the term @xmath182 is the product of @xmath53 s and @xmath183 remaining after all the powers of @xmath54 and @xmath55 have been pushed to the right , namely @xmath184\prod_{a=1}^b \big(\prod_{k < i_a}v^{-\epsilon_kl_k.2\delta_a}\big ) k^{2\delta_a}\>.\ ] ] where by abuse of notation we have written @xmath185 and @xmath186 in place of @xmath187 and @xmath188 respectively . collecting terms and noting that @xmath189 , @xmath190 since both @xmath182 and @xmath180 preserve the weight , @xmath191 where the superscripts @xmath82 denote the @xmath169 matrix element in the evaluation on @xmath83 . thus @xmath192 is obtained from the above expression by replacing all occurrences of @xmath53 and @xmath16 by @xmath94 and @xmath95 , respectively . we may now carry out the sum over @xmath181 in ( [ jonesqg ] ) for @xmath147 , since the sum on @xmath193 may be taken out to give a factor ( where @xmath194 ) @xmath195_{00 } \kern-14pt = v^{{\sigma_j\over2}(\lambda+2\sum\limits_{k < j+}\epsilon_k{}l_k ) ( \lambda+2\sum\limits_{k < j-}\epsilon_k{}l_k)}\ ] ] moving to the zero framed knot introduces a framing correction of @xmath196 , that is @xmath197 . the contribution of the term @xmath192 to @xmath198 becomes simply a power of @xmath76 . substituting into ( [ jonesqg ] ) , we get the zero framed colored jones polynomial as @xmath199 where ( again @xmath194 ) @xmath200 the function @xmath201 is a linear function of @xmath82 , and depends quadratically on the parameters @xmath202 . looking at only the parity of terms ( that is , mapping onto @xmath203)$ ] , @xmath204 since @xmath205 ( each term @xmath202 appears twice with opposite signs ) , the first term is @xmath206 . reversing the order of summation in the last term , @xmath207 , where @xmath208 is the number of segments on the gauss diagram between the two occurrences of the crossing @xmath166 , that is between @xmath209 and @xmath210 . observing that @xmath211 is always even and that @xmath208 is always odd ( as follow from changing the signs of crossings in the knot diagram so as to obtain an unknot and then applying reidemeister moves to untie it and observing that in the process these parities are preserved ) , it follows that @xmath212 the final evaluation required in order to remove all traces of the quantum group @xmath62 from our formula for @xmath198 is that of @xmath213 in ( [ bl ] ) , namely the @xmath169 matrix element in the representation @xmath83 of the product of terms @xmath214 and @xmath215 described by the gauss diagram . using the explicit form for the basis for @xmath83 , one can read the product @xmath180 from right to left , each term @xmath54 pushes up the level , while each @xmath55 pushes it down with a factor @xmath216[\mu - i]$ ] starting at level @xmath67 . reading instead from left to right , @xmath54 decreases the level while @xmath55 raises the level and introduces the factor @xmath217[\mu-1-i]$ ] when starting from level @xmath67 . thus we obtain @xmath218[\mu-1-i]\ ] ] where @xmath219 is the level at the start of the term @xmath220 contributed by @xmath210 in the gauss diagram ( if any @xmath221 is negative , the expression vanishes ) . writing @xmath217[\mu-1-i]=v^{-\sigma\mu}\{i+1\}_\sigma\{\mu-1-i\}_\sigma$ ] , we obtain @xmath222 substituting into ( [ jonesq ] ) and putting @xmath223 $ ] , we obtain the following result.2ex * theorem 1 @xcite*_the colored jones function of a zero framed knot is given by @xmath224 where @xmath225 and @xmath226 is a function dependent on the gauss diagram , linear in @xmath82 and quadratic in @xmath202 , with integer coefficients , given explicitly by ( [ dl ] ) and ( [ cl ] ) below._2ex several remarks and standard facts about the colored jones function follow immediately from this theorem . * in @xcite , this theorem is only stated explicitly for braid closures . a very similar ( though inequivalent ) state sum for the colored jones function was recently obtained in theorem 7 , @xcite . * considering @xmath227 , each term can be thought of as a power series in @xmath228 whose coefficients are polynomials in @xmath82 . the contribution from a particular set of @xmath153 non - negative integers @xmath202 is divisible by @xmath229 . that is , the coefficient of @xmath230 comes only from @xmath202 s with @xmath231 , a finite number of contributions . so @xmath198 is meaningful as a formal power series in @xmath228 whose coefficients are polynomials in @xmath82 ( even when @xmath82 is not an integer ) ; this is the jones function . * if @xmath81 , then the contribution from any term for which @xmath232 will vanish . thus only a finite number of @xmath233 s will contribute , giving a polynomial ; this is the colored jones polynomial . * for each @xmath234 , the contribution is a polynomial in @xmath13 and @xmath235 , so that @xmath236 can be written as a power series in @xmath228 whose coefficients are polynomials in @xmath235 . ( see the melvin - morton - rozansky conjecture of @xcite , @xcite proved in @xcite . ) * since @xmath237 $ ] , it follows that the contribution from @xmath234 is divisible by @xmath238 . at a root of unity of order @xmath16 , only terms with @xmath239 @xmath240 will contribute . again there are a finite number of contributions . * the coefficient of @xmath82 in @xmath226 simplifies to @xmath241 where @xmath242 . when the knot is presented in a blackboard framing which is also zero framed , then @xmath243 , so that @xmath244 where now @xmath245 has the meaning of half the sum of the signs of crossings encountered when tracing around the knot from the overcrossing arc of the @xmath246 crossing to its undercrossing arc ; equivalently , it is the linking number of the two parts of the knot which the @xmath246 crossing , if split so as to give a link , separates . * in the limit @xmath247 with @xmath235 ( or @xmath248 ) constant , we obtain a formal power series in @xmath249 . in particular , the kashaev invariant obtained by looking at the colored jones polynomial in color @xmath16 at the @xmath14 root of unity @xmath250 is obtained by setting @xmath251 , @xmath252 , @xmath253 where @xmath254 is the constant term in @xmath255 , namely @xmath256 * when @xmath257 , the colored jones polynomial reduces to the one - variable jones polynomial @xcite . it can be checked that the only non - zero contributions to the sum in theorem 1 in this case come from @xmath258 for which @xmath259 for all @xmath67 . this sum is similar to ( though different from ) the beautifully simple state model given by kauffman for the bracket polynomial in @xcite . put concisely , the colored jones function depends on two parameters , @xmath13 and @xmath82 , and one can make sense of it if at least one of the following holds : ( i ) @xmath13 is a root of unity ; ( ii ) @xmath81 ; ( iii ) @xmath227 is considered as a formal parameter with expressions as formal power series in @xmath228 and @xmath82 is fixed ; ( iv ) @xmath13 is considered as a formal parameter with expressions as formal power series in @xmath228 and @xmath260 is fixed . suppose that @xmath13 is a @xmath14 root of unity . the prescription of @xcite defines for a link @xmath5 with @xmath153 components and framing @xmath263 on the @xmath264 component , @xmath265 ^ 2q^{\sum_if_i\omega(\mu_i)}\>,\ ] ] where @xmath266 is the framing normalization , that is @xmath267 . the @xmath261 invariant of a 3-manifold @xmath0 obtained from @xmath6 by surgery around a link @xmath5 is then @xmath268 where @xmath269 are the numbers of positive / negative eigenvalues of the linking matrix of @xmath5 and @xmath270 are the bracket values for the unknot with framings @xmath271 . when @xmath5 is a knot ( @xmath272 ) with framing @xmath273 , so that @xmath0 is an integer homology 3-sphere , this reduces to @xmath274 ^ 2 j'_{l_0}(\mu ) \over\sum\limits_{\mu=1}^{k-1}q^{f\mu^2/4}[\mu]^2}\>.\ ] ] the @xmath262 invariant is obtained by the same procedure except that the sums are all restricted to @xmath275 even ; see @xcite . following le and habiro , we now show how to obtain habiro s cyclotomic form for the ohtsuki series from theorem 1 . assume that @xmath0 is obtained by @xmath273-surgery around a knot @xmath5 , so that @xmath0 is an integer homology sphere . the expression for @xmath198 in theorem 1 is a sum over @xmath234 of terms which are polynomial in @xmath276 and @xmath13 ( and their inverses ) . a term @xmath277 in @xmath198 contributes @xmath278 ^ 2q^{m\mu } \over\sum\limits_{\mu=1}^{k-1}q^{f\mu^2/4}[\mu]^2}\ ] ] to @xmath15 . both summands vanish at @xmath279 and are even in @xmath82 , so that the sum can be replaced by a complete period @xmath280 . completing the square , @xmath281 , thus the contribution to @xmath15 of a term @xmath277 in @xmath282 is @xmath283\>.\ ] ] let @xmath284 be the ring of formal power series in @xmath249 whose coefficients are polynomials in @xmath285 with integer coefficients . for @xmath273 , define a map @xmath286 $ ] linear over @xmath287 $ ] by @xmath288 so that @xmath289 . 2ex*lemma 1*_for any @xmath290 , @xmath291 , @xmath292 is divisible by the @xmath293 cyclotomic polynomial @xmath294 in @xmath295$]._2ex * proof * suppose @xmath296 is a root of unity of order @xmath234 . it is sufficient to show that the expression in the lemma vanishes when evaluated at any such @xmath297 . expanding out the product , @xmath298 where @xmath299 and @xmath300 denote the sum of elements and the number of elements in the set @xmath301 , respectively . under the map @xmath302 , which adds 1 ( modulo @xmath234 ) to every element of @xmath301 , the number @xmath299 increments by @xmath300 while @xmath300 stays unchanged . under repeated application of this map , we get orbits of size @xmath234 ( for @xmath303 ) and two orbits with one element , namely the empty set and the whole set . the contribution from any orbit of size @xmath234 contains a factor @xmath304 which vanishes when evaluated at @xmath297 . the contribution from the two size one orbits is @xmath305 when evaluated at @xmath297 , @xmath306 while @xmath307 and @xmath308 , and again the contribution vanishes.@xmath309 now for any @xmath310 , @xmath311 can be written as a linear combination ( over @xmath295 $ ] ) of expressions of a similar form with chains of length @xmath312 , @xmath313 . as a corollary , the expression in lemma 1 is divisible by @xmath314 for any @xmath310 and therefore also ( since they are coprime ) by @xmath315 . here $ ] is the minimal monic polynomial in @xmath13 which vanishes at all roots of unity of order @xmath317 ( but not at @xmath318 ) . 2ex*corollary*_for any @xmath290 , @xmath291 , @xmath292 is divisible by @xmath319 in @xmath295$]._2ex 2ex*lemma 2*_for any integers @xmath118 , @xmath320 , @xmath321 is divisible by @xmath322 . _ 2ex * proof * let @xmath323 $ ] be defined by @xmath324 for @xmath325 , extended linearly over @xmath287 $ ] . write @xmath326 and extend @xmath327 linearly to formal power series with rational coefficients . then @xmath328 . comparing coefficients of powers of @xmath329 , @xmath330 while @xmath331 . in particular , for @xmath332 , @xmath333 is divisible by @xmath334 . now @xmath335 and @xmath327 are related by @xmath336 expanding as a power series in @xmath228 , whose coefficients are polynomials in @xmath82 , @xmath337 is divisible by @xmath338 , and such that the coefficient of @xmath339 ( @xmath340 ) is a polynomial in @xmath82 of degree at most @xmath341 . replacing @xmath228 by @xmath342 preserves these properties ( though coefficients of powers of @xmath342 may now be polynomials with non - integral , but rational , coefficients ) . hence applying @xmath327 we obtain an element of @xmath343 $ ] divisible by @xmath344 . the lemma now follows . @xmath309 2ex*lemma 3*_for any integers @xmath118 , @xmath320 , @xmath202 , @xmath345 is divisible by @xmath346 in @xmath287 $ ] and in @xmath295 $ ] . _ 2ex * proof * note that @xmath347 and @xmath335 interchange when @xmath13 and @xmath47 are interchanged , so it is sufficient to prove the lemma for @xmath348 . next observe that the expression in the lemma is actually a polynomial in @xmath31 and so divisibility in @xmath287 $ ] and in @xmath295 $ ] are equivalent . since @xmath228 and @xmath349 are coprime in @xmath287 $ ] , it is sufficient to prove divisibility by @xmath350 and by @xmath351 separately ; for the first we work in @xmath287 $ ] ( lemma 2 ) and the second in @xmath295 $ ] ( corollary).@xmath309 putting lemma 3 together with theorem 1 from the last section gives the following . * theorem 2*_the wrt invariant of a manifold , @xmath0 , obtained by @xmath273 surgery around a knot , @xmath5 , is given by @xmath352 in which the contribution from @xmath233 is divisible by @xmath346 . here @xmath353 while @xmath254 and @xmath354 are given by ( [ cl ] ) and ( [ dl ] ) respectively . this is meaningful when @xmath13 is a root of unity ( then the sum is finite as the only non - vanishing contributions are when @xmath239 for all @xmath166 ; the sum is the wrt invariant ) and as a formal power series in @xmath228 ( the sum is the ohtsuki series , @xmath28)._2ex in @xcite , habiro obtains a more symmetric cyclotomic form for ohtsuki series which is needed to prove integrality in general . for the case of surgery around knots , the above simple - minded approach is sufficient . suppose that @xmath0 is an integral homology sphere obtained by surgery around a knot . then the ohtsuki series @xmath28 is a formal power series @xmath355 in @xmath249 . in this section we use the formula given in theorem 3 to determine a bound on the growth rate of coefficients @xmath356 with @xmath357 . by the theorem , the only terms @xmath233 which contribute to @xmath358 are those for which @xmath359 , that is @xmath360 ; there are @xmath361 such @xmath233 s . the double product in the summand in theorem 2 is a product of @xmath362 factors of the form @xmath363 where @xmath364 . thus the contribution from @xmath233 to @xmath28 can be expanded out to give a sum of @xmath365 terms of the form @xmath366 where @xmath367 and @xmath368 , while @xmath369 . using ( [ dl ] ) and ( [ cl ] ) , one finds estimates @xmath370 and @xmath371 , so that the order of the coefficient of @xmath372 ( @xmath373 ) in @xmath374 is that of @xmath375 ( @xmath118 is a constant ) . the final term to estimate is of the coefficient of @xmath376 in the product of @xmath153 @xmath13-binomial coefficients on the right hand side of ( [ zinfterm ] ) . each of these @xmath13-binomial coefficients is of the form @xmath377 with @xmath378 ; such a coefficient is a sum of @xmath379 powers of @xmath13 ( with exponents at most @xmath380 ) and so the coefficient of a power @xmath339 is at most @xmath381 taking a product of @xmath153 such terms can accumulate in the coefficient of @xmath339 at most @xmath382 and thus the coefficient of @xmath230 in ( [ zinfterm ] ) is at most @xmath383 putting all the above estimates together gives the order of @xmath358 to be at most that of @xmath384 which is an order of growth of the bound for which @xmath385 asymptotically grows linearly with @xmath386 , with slope of order @xmath387 . * theorem 3*_there exist bounds @xmath388 dependent only on @xmath357 and the number of crossings @xmath153 in the knot @xmath16 , for which @xmath389 for all @xmath390 while @xmath391 exists and is of order @xmath387._2ex this should be compared with the slope conjecture of @xcite which states that @xmath392 is asymptotically linear in @xmath386 , with coefficient , @xmath393 , called the slope of @xmath0 . combining with the above result , it is reasonable to conjecture that @xmath393 is of order @xmath387 , when @xmath0 is obtained by surgery around a knot with @xmath153 crossings . this is consistent with known results ( @xcite , @xcite ) on the slope for seifert fibred manifolds . in the literature ( see @xcite , @xcite , @xcite , @xcite , @xcite ) there are computations of the colored jones function for the simplest knots , the trefoil , figure-8 knot and more generally , torus knots and twist knots . applying theorems 1 and 2 , we reproduce the simplest of these results , although some are in different forms . suppose that @xmath5 is the right hand trefoil , presented as in figure 6 . in general , the formulae obtained will be simpler if the basepoint is chosen so that the first crossing encountered is an overcrossing ( and the last before returning to the basepoint is an undercrossing . crossing 3 is a left pointing crossing and the appropriate enhanced gauss diagram is shown on the right in figure 6 . in this case @xmath394 and the six points ( from crossings ) on the outer circle of the gauss diagram are marked @xmath395 , @xmath396 , @xmath397 , @xmath398 , @xmath399 and @xmath400 , respectively . also there is just one blob @xmath137 , so that @xmath401 and @xmath402 . the signs of the crossings @xmath155 are all @xmath403 . hence @xmath404 non - vanishing contributions to @xmath282 appear only when @xmath221 are all non - negative , so that @xmath405 . then @xmath406 while @xmath407 . we calculate @xmath408 so that @xmath409 and so @xmath410 . theorem 1 now gives the jones function of the trefoil as @xmath411 while the kashaev invariant is @xmath412 these formulae may be put in more standard form as @xmath413 where @xmath414 . this formula was obtained by habiro@xcite and le , as well as another one , @xmath415 the figure 8 knot ( @xmath416 ) is naturally presented in blackboard framing as a zero framed knot . figure 5 shows its enhanced gauss diagram along with the labeling of points on the circle . here @xmath417 with signs @xmath418 and @xmath419 . there are two blobs @xmath137 so that @xmath420 , @xmath421 . similar to the case of the trefoil , the only non - vanishing contributions come from @xmath405 . set @xmath422 and @xmath423 . then @xmath424 and @xmath407 . also @xmath425 , @xmath426 so that from ( [ dl ] ) @xmath427 the last term in ( [ cl ] ) requires an enumeration of all triples @xmath428 for which @xmath429 and @xmath430 ; the part of the list for which @xmath431 is @xmath432 , @xmath433 , @xmath434 , @xmath435 , @xmath436 , @xmath437 , @xmath438 , @xmath439 , @xmath440 and @xmath441 . substituting into ( [ cl ] ) gives @xmath442 theorem 1 supplies the jones function of the knot @xmath416 as @xmath443 where @xmath444 . the kashaev invariant is then @xmath445 this is to be compared with habiro @xcite and le s formulae @xmath446 in which the symmetry @xmath447 due to the amphichirality of @xmath416 is apparent . 2ex * acknowledgements * the first author would like to thank thang le and lev rozansky for many conversations on the subject of computations of 3-manifold invariants . the second author would like to acknowledge partial support from an excellence grant from the hebrew university .
we give a self - contained treatment of le and habiro s approach to the jones function of a knot and habiro s cyclotomic form of the ohtsuki invariant for manifolds obtained by surgery around a knot . on the way we reproduce a state sum formula of garoufalidis and le for the colored jones function of a knot . as a corollary , we obtain bounds on the growth of coefficients in the ohtsuki series for manifolds obtained by surgery around a knot , which support the slope conjecture of jacoby and the first author . = cmr7 = eufm10 # 1#2 \ { cc-4pt#1 - 4pt + -4pt#2 - 4pt + } # 1#2 # 1#2 ( cc-4pt#1 - 4pt + -4pt#2 - 4pt + ) # 1max#1 # 1#2(#1#2 ) * on habiro s cyclotomic expansions of the ohtsuki invariant * 2ex ruth lawrence and ofer ron _ einstein institute of mathematics , hebrew university of jerusalem _ _ givat ram , 91904 jerusalem , israel _ _ e - mail address : _
the calculation of gravitational radiation produced in the merger of two black holes is of considerable theoretical interest . it may also provide templates of wave - forms for analysis of ligo - virgo data . to be of practical interest to matched filtering , the calculational methods will have to recover many wave - periods over an extended parameter range , such as in the problem of binary black hole coalescence with a range of black hole masses and spins . it poses the challenge of efficient and stable phase - accurate numerical methods . it has been appreciated that this requires an inherently stable formulation of the dynamical evolution of spacetime . spectral methods provide the most efficient representation of functions that are periodic and everywhere analytic , and preserve accurate phase information in hyperbolic evolution . here , we consider the 3 + 1 hamiltonian equations for the dynamical behavior of spacetime @xcite . they are attractive , in describing the evolution in terms of quantities that have unambiguous geometric interpretations . however , for arbitrary choices of gauge , they are known to give rise to computational instabilities , which have thusfar prevented their application to large - scale numerical relativity . here , we propose a new gauge for time - evolution , which describes a correlation between the lapse function and the extrinsic curvature of the foliation of three - surfaces in spacetime . linearized stability analysis shows that the full system of hamiltonian equations now becomes hyperbolic with respect to arbitrary perturbations of the three - metric . next , we focus on black hole spacetimes which are asymptotically quiescent @xcite . this asymptotic property , if present in the initial data , is preserved for all time by causality in the equations of general relativity . it can be exploited by compactification in spherical coordinates with complex radius , @xmath3 with complex radius @xmath4 , @xmath5 , where @xmath6 denotes a dimensional continuation parameter for a spacetime with total mass energy @xmath7 . here , @xmath8 can be chosen sufficiently large , i.e. , at least on the order of @xmath7 , to avoid singularities associated with black holes , even those that may carry spin . in this approach , we focus on the _ outer _ expansion of the green s function . it makes possible a three - dimensional spectral representations after application of a mbius transformation ( which preserves conformally flat initial data ) . the combination of hyperbolic hamiltonian equations and spectral representations offers a novel starting point for studying the dynamical evolution of black hole spacetimes . for illustrative purposes , we consider the conformally flat initial data of schwarzschild black holes , as in the problem of a head - on collision of two black holes . here , conformally flat data are special , in representing turnings points corresponding to the up or down position of a pendulum . the proposed lapse function is natural choice about such turning points . to illustrate compactification by a mbius transformation , we explicitly identify newton s law in terms of the evolution of the quadrupole deformations in area of poloidal flux surfaces , given by the moment of inertia @xmath9 of spacetime . at large distances from the source region , the two wave - modes in the gravitational field appear as planar waves in the three - metric @xmath10 , commonly expressed in the transverse traceless gauge as @xmath11 @xmath12 denote the amplitude of @xmath13 and @xmath14 plane - waves along the @xmath15direction with wave - number @xmath16 . the three - dimensional ricci tensor associated with @xmath10 hereby assumes the following structure . for the @xmath13 waves with @xmath17 , we have @xmath18 where @xmath19 @xmath20 and @xmath21 , and for the @xmath14 waves with @xmath22 , @xmath23 , @xmath24 , we find , similarly , @xmath25 where @xmath26 . hence , we have the following exact algebraic identities : @xmath27 where the numerical indices refer to normalized directions in the coordinate directions @xmath28 . the identies ( [ eqn_id ] ) show that the two polarization modes satisfy traveling wave solutions @xmath29 in the large - distance approximation ( neglecting @xmath30 curvature terms , in this section ) with vanishing lapse functions @xmath31 and constant lapse function @xmath32 , satisfying @xmath33 where @xmath34 using the numerical indices to denote the normalized directions in @xmath35 ) . the angular directions 22 - 33 " and 23 " therefore represent the _ hyperbolic directions _ , which contain all wave - motion . we consider the 3 + 1 decomposition of the line - element in terms of @xmath10 @xcite @xmath36 the lapse and shift functions can be seen to define the motion of observers with velocity four - vector @xmath37 , for example , by considering their acceleration @xmath38 which is determined entirely by the redshift factor @xmath39 when @xmath40 . in general , the lapse and shift functions appear in the conservation of energy and momentum , of test particles moving along geodesics . we need not impose the time - symmetric gauge condition with @xmath41 . the three - metric @xmath10 satisfies the hamiltonian evolution equations @xmath42 where @xmath43 and @xmath44 denotes the gauged three - tensor @xmath45 we propose the time - evolution accoring to @xmath46 where @xmath47 and @xmath48 . the gauge condition @xmath0 is _ curvature - driven evolution_. it is different from the product of curvature and lapse function in the harmonic slicing condition @xmath49 in eqs . ( 69)-(77 ) of @xcite ; see also @xcite for a recent review . in the asymptotically flat region about @xmath50 and @xmath51 , we have @xmath52 we recall that @xcite @xmath53 where @xmath54 , where @xmath55 refers to the trace of the metric perturbations . for harmonic perturbations of the for @xmath56 , we have , with conservation of momentum , @xmath57 , @xmath58 , so that @xmath59 and hence @xmath60 we conclude that @xmath61 gives rise to the dispersion relation @xmath62 for arbitrary , small amplitude metric perturbations . we conclude that the system @xmath1 in ( [ eqn_k ] ) is asymptotically stable by virtue of the gauge choice @xmath0 . this contrasts with the more common derivation of ( [ eqn_d ] ) in the so - called transverse traceless gauge , or harmonic coordinates neither of these two coordinate conditions are used here . the conformal decomposition of the ricci tensor for @xmath63 satisfies @xcite @xmath64 + m\phi^{-2}[(1+m)(n-2){d}_i\phi{d}_j\phi + ( 1-m*(n-2)){g}_{ij } { d}^p\phi { d}_p\phi].$ ] for @xmath65 and @xmath66 , it reduces to @xmath67 + 2\phi^{-2}[3 { d}_i\phi{d}_j\phi - { g}_{ij } { d}^p\phi { d}_p\phi],$ ] giving the familiar result @xmath68 . \label{eqn_cid}\end{aligned}\ ] ] ( solid curved line ) or equivalently repulsive , following closure over infinity in the domain of convergence of the outer expansion of the green s function , i.e. , in the space of the transformed variable @xmath69 ( dotted curved lines ) . for the latter , newton s law represents the leading order evolution of the area of the poloidal surface elements @xmath70 , governed by the moment of inertia associated with the separation @xmath71 between the two black holes.,width=302,height=302 ] time - symmetric initial data around one or multiple non - rotating black holes can be conveniently described in a conformally flat representation @xcite on the basis of ( [ eqn_cid ] ) . it gives rise to solutions for vacuum spacetimes in terms of the green s function @xmath72 of flat spacetime , where @xmath73 denotes the position of a point source . spacetime of a single black hole is given by the conformal factor @xmath74 where @xmath75 , @xmath76 and the expansion refers to the _ outer expansion _ of @xmath77 in complex spherical coordinates @xmath78 , @xmath79 , by exploiting analyticity at infinity , i.e. , asymptotic quiescence . the same construction applies to two black holes , @xmath80 where @xmath81 and @xmath82 . without loss of generality , we may choose @xmath83 . according to the hamiltonian evolution equations , the associated shift function , for a static foliation of spacetime , i.e. , @xmath84 , satisfies @xmath85 for the conformally flat data at hand , it reduces to @xmath86 , or @xmath87 . the leading - order structure of spacetime can be studied by considering poloidal surface elements @xmath88 in the outer region @xmath70 . a measure for the @xmath2 deformations in surface elements is @xmath89 in @xmath90 , where @xmath91 refers to time - independent divergences . it gives a geometric equivalence about @xmath92 @xmath93 to the moment of intertia . the gauged tensor @xmath94 has a regular taylor series expansion about infinity , for the given asymptotically quiescent initial data . for the equal mass , two - hole solution with total mass - energy @xmath95 , moment of intertia @xmath96 , and symmetric position about the origin , the conformal factor @xmath97 we find the large-@xmath4 asymptotics @xmath98/(1-x^2 ) & 0\\0 & 0 & -z^{2}p_0(1-x^2 ) \end{array } \right ) , \label{eqn_wii}\end{aligned}\ ] ] satisfying @xmath99 with @xmath100 the time - evolution of the @xmath2 deformation of poloidal surface elements is given by @xmath101 where @xmath102 , where @xmath103 for the time - symmetric initial data , the quadratic off - diagonal terms make no initial contribution , leaving @xmath104 at @xmath105 , the integrand reduces in the gauge @xmath106 with the additional time - symmetric gauge condition @xmath41 at @xmath105 ( a newtonian gauge choice , as in a turning point ) to @xmath107 for completely time - symmetric initial data , we thus recover newton s law for two equal mass black holes of mass @xmath108 and separation @xmath109 in @xmath110 , we find @xmath111 for initial evolution of @xmath112 at the maximal extent @xmath113 ( the lower limit of @xmath4 ) of the poloidal surface elements ( in the outer expansion considered ) . thus , for a fixed @xmath70 , it takes @xmath114 to relax @xmath115 by the leading - order @xmath116 forcing terms in ( [ eqn_wii ] ) , when applied to the asymptotic equations . the asymptotics @xmath115 for large @xmath4 hereby persists for all finite time . this observation is consistent with the a priori notion of preserving asymptotic quiescence . here , newton s law ( [ eqn_new ] ) associated with the deformations of the poloidal surface elements @xmath70 represent the underlying non - local interaction . we have shown that the 3 + 1 hamiltonian equations become hyperbolic , when the evolution of the lapse function is driven by the extrinsic curvature tensor . the resulting hyperbolic system forms a starting point for three - dimensional spectral methods in numerical relativity , after compactification by a mbius transformation for spacetimes that satisfy asymptotic quiescence , wherein newton s law arises in the dynamics of the quadrupole deformations of poloidal surface elements . future work will focus on some numerical examples . * acknowledgment . * the author gratefully acknowledges stimulating discussions with a. spallicci , m. volkov , g. barles , and members of the fdration denis poisson . this work is supported , in part , by le studium ias of the universit dorlans . abrahams , a. , anderson , a. , choquet - bruhat , y. & york jr . , j.w . , 1997 , class . quantum grav . , a9 arnowitt , r. , deser , r. , & misner , c.w . , 1962 , in gravitation : an introduction to current research , edited by l. witten ( wiley , new york ) , pp.227 brill , d.r . , & lindquist , r.w . , 1963 , d. , 131 , 471 brown , j.d . , 2008 , gr - qc/0803.0334v2 lindquist , r.w . , 1963 , phys . rev . , 4 , 938 jansen , n. , diener , p. , khokhlov , a. , & novikov , i. , 2003 , class . grav . , 20 , 51j thorne , k.s . , price , & mcdonald , _ the membrane paradigm _ van putten , m.h.p.m . , 2006 , proc . acad . sc . , 516 wald , r.m . , 1984 , _ general relativity _ chicago press , chicago )
the 3 + 1 hamiltonian formulation in the gauge @xmath0 on the lapse function fixes the direction of time associated with the trace @xmath1 of the extrinsic curvature tensor . the hamiltonian equations hereby become hyperbolic . we study this new system for black hole spacetimes that are asymptotically quiescent , which introduces analyticity properties that can be exploited for numerical calculations by compactification in spherical coordinates with complex radius following a mbius transformation . conformal flat initial data of two black holes are hereby invariant , and correspond to a turn point in a pendulum , up for a pair of separated black holes and down for a single black hole . here , newton s law appears in the relaxation of @xmath2 deformations of semi - infinite poloidal surface elements , defined by the moment of inertia of the binary .
the history of the investigations of half - metallic ferromagnets ( hmf ) starts from the electronic structure calculation for nimnsb @xcite ; later a number of other examples were discovered , e.g. , cro@xmath0 , fe@xmath1o@xmath2 , a number of the heusler alloys co@xmath0mnz and rmnsb ( for a review , see refs . ) . these substances have metallic electronic structure for one spin projection ( majority- or minority - spin states ) , but for the opposite spin direction the fermi level lies in the energy gap . owing to this fact hmf attract now a growing attention of researchers in connection with `` spintronics '' , or spin - dependent electronics @xcite . the spin - up and spin - down contributions to electronic transport properties have different orders of magnitude , which can result in a huge magnetoresistance for heterostructures containing hmf @xcite . some evidences of the hmf behavior in colossal magnetoresistance ( cmr ) materials like la@xmath3sr@xmath4mno@xmath1 found by using tunneling spectroscopy @xcite and photoemission technique @xcite have increased considerably the interest in the half - metallic ferromagnetism ; however , the situation in the cmr systems is controversial , as demonstrate andreev reflection experiments @xcite . peculiar band structure of hmf results in an important role of incoherent ( non - quasiparticle , nqp ) states which occur because of correlation effects @xcite . the appearance of nqp states in the energy gap near the fermi level @xcite is one of the most interesting correlation effects typical for hmf . the origin of these states is connected with `` spin - polaron '' processes : the spin - down low - energy electron excitations , which are forbidden for hmf in the one - particle picture , turn out to be possible as superpositions of spin - up electron excitations and virtual magnons . the density of these states vanishes at the fermi level @xmath5 for temperature @xmath6 , but increases drastically at the energy scale of the order of a characteristic magnon frequency @xmath7 . the existence of nqp states is relevant for spin - polarized electron spectroscopy @xcite , nmr @xcite , core - level spectra of the hmf @xcite , and subgap transport in ferromagnet - superconductor junctions ( andreev reflection ) @xcite . several experiments could be performed in order to clarify the impact of the nqp states on spintronics . in particular , @xmath8 characteristics of half - metallic tunnel junctions for the case of antiparallel spins are completely determined by nqp states @xcite . recently the density of nqp states has been calculated from first principles for a prototype hmf , nimnsb @xcite , and for cras @xcite . on the other hand , hmf are very interesting conceptually as a class of materials which may be convenient to treat many - body solid state physics that is essentially beyond band theory . it is accepted that usually many - body effects lead only to renormalization of the quasiparticle parameters in the sense of landau s fermi liquid ( fl ) theory , the electronic _ liquid _ being _ qualitatively _ similar to the electron _ gas _ ( see , e.g. , refs . ) . on the other hand , nqp states in hmf are not described by the fl theory . as an example of highly unusual properties of the nqp states , we note that they can contribute to the @xmath9-linear term in the electron heat capacity @xcite , despite their density at @xmath5 is zero for @xmath6 . spin - polarized scanning tunneling microscopy ( stm ) @xcite is a very efficient new method which enables one to probe directly the spectral density with spin resolution in magnetic systems . the spin - polarized stm should be able to probe the nqp states via their contribution to the differential tunneling conductivity @xmath10 . at zero temperature , nqp states arise only above @xmath5 for the case of minority - spin gap and only below @xmath5 for the majority - gap hmf @xcite . unlike the photoemission spectroscopy which probes only occupied electron states , stm detects the states both above and below @xmath5 , depending on the sign of bias . theoretical investigation of nqp contributions to stm spectra is the aim of the present paper . the paper is organized as follows . in sect.2 we discuss a general expression for the tunneling current @xmath11 as applied to hmf . in sect.3 the effect of surface potential and other spatial inhomogeneities on the nqp spectral density is considered . in sects.4 and 5 we calculate the energy and temperature dependences of @xmath10 and treat the problem of tunneling - current spin - polarization at finite temperatures . in sect.6 the bias dependences of the tunneling conductance are discussed . a general expression for the tunneling current in the lowest order in the tunneling matrix elements has the form @xcite @xmath12 \label{curr}\ ] ] where @xmath13 is electron charge , @xmath14 is the spin projection , @xmath15 is the bias , @xmath16 is the fermi distribution function , greek ( latin ) indices label electron eigenstates for the sample ( tip ) @xmath17 @xmath18 is the current matrix element , @xmath19 is the free electron mass ( the surface integral in eq.([matelem ] ) is taken over arbitary area between the tip and the sample ) , and @xmath20 is the corresponding spectral density , @xmath21 is the sample green s function , @xmath22 @xmath23 being the creation operators for conduction electrons . it is worthwhile to emphasize that the expression ( [ curr ] ) takes into account correlation effects in a formally exact way , assuming that the tunneling probability is small . in fact , the latter condition should be satisfied for proper stm measurements , otherwise they can not be considered as a true probe . in the wkb approximation eq.([curr ] ) takes the form @xcite @xmath24 g_{s}^{\sigma } \left ( e\right ) \label{tok}\ ] ] where @xmath25 is the average of sample and tip work functions ( which is assumed to be large in comparison with @xmath26 for simplicity ) , @xmath27 is the distance between the surface and the tip . here @xmath28 is the density of states ( dos ) of the tip material , which is supposed to be smooth and thus is replaced by its value at the fermi energy , and @xmath29 is the density of states of the sample with zero in - plane component of the wave - vector : @xmath30 so that the summation is performed only over two points of the fermi surface . this condition means that the tunneling probability has a sharp ( at not too small @xmath27 ) maximum for the states with velocity direction normal to the surface . for a generic multi - sheet fermi surfaces the condition of the dominant tunneling is , generally speaking , more complicated , but this modifies only some weakly bias - dependent factors . since stm probes only surface one has to discuss first modification of nqp states in comparison with the case of ideal bulk crystal . the existence of nqp states at the surface of hmf has been demonstrated for a narrow - band hubbard model @xcite . here we present a general derivation valid in the case of arbitrary inhomogeneity . to describe the effects of electron - magnon interaction for the inhomogeneous case we use the formalism of the exact eigenfunctions developed earlier for the impurity - state problem in a ferromagnetic semiconductor @xcite . the corresponding hamiltonian of the @xmath31 exchange model reads @xmath32 where @xmath33 is the potential energy ( with account of the electron - electron interaction in the mean field approximation ) which is supposed to be spin dependent , @xmath34 is the field operator for the spin projection @xmath35 @xmath36 is the spin density of the localized - moment system , @xmath37 is its fluctuating part , the effect of the average spin polarization @xmath38 being included into @xmath33 . we use an approximation of contact electron - magnon interaction described by the @xmath31 exchange parameter @xmath39 , @xmath40 is the heisenberg hamiltonian of localized spin ( for simplicity we neglect the inhomogeneity effects for the magnon subsystem ) . passing to the representation of the exact eigenfunctions of the hamiltonian @xmath41@xmath42 one can rewrite the hamiltonian ( [ hamilt ] ) in the following form : @xmath43 where @xmath44 we take into account again the electron - spectrum spin splitting in the mean - field approximation by keeping the dependence of the eigenfunctions on the spin projection . we restrict ourselves to the spin - wave region where we can use for the spin operators the magnon ( e.g. , dyson - maleev ) representation . then we have for the one - electron green s function @xmath45 ^{-1 } , \label{dys}\ ] ] with the self - energy @xmath46 describing correlation effects . we start with the perturbation expansion in the electron - magnon interaction . to second order in @xmath39 one has @xmath47 with @xmath48 where @xmath49 , @xmath50 is the magnon energy , @xmath51 is the bose function . using the expansion of the dyson equation ( [ dys ] ) we obtain for the spectral density @xmath52 the second term in the right - hand side of eq.([n(e ) ] ) gives the shift of quasiparticle energies . the third term , which arises from the branch cut of the self - energy , describes the incoherent ( non - quasiparticle ) contribution owing to scattering by magnons . one can see that this does not vanish in the energy region , corresponding to the `` alien '' spin subband with the opposite projection @xmath53 . neglecting temporarily in eq.([q ] ) the magnon energy @xmath54 in comparison with typical electron energies and using the identities @xmath55 we derive @xmath56 \left\langle \nu \downarrow \right| \delta \left ( e - e^{\prime } -% \mathcal{h}_{0}^{\uparrow } \right ) \left| \nu \downarrow \right\rangle \label{sigmadn}\end{aligned}\ ] ] here we restrict ourselves only to the case of zero temperature @xmath6 neglecting the magnon occupation numbers @xmath57 . using the tight - binding model for the ideal - crystal hamiltonian we find in the real - space representation @xmath58 \left ( -\frac{1}{\pi } \text{\textrm{im}}g_{% \mathbf{r , r}}^{\uparrow } ( e^{\prime } ) \right ) \delta _ { \mathbf{r , r}^{\prime } } \label{sig_dn}\end{aligned}\ ] ] where @xmath59 are lattice site indices , and therefore @xmath60 following the method developed by us earlier @xcite one can generalize the above results to the case of arbitrary @xmath31 exchange parameter . simplifying the sequence of equations of motion ( cf . ref . ) we obtain the integral equation @xmath61 where @xmath62 note that the equation ( [ int ] ) is exact in the case of empty conduction band ( @xmath63 one current carrier , ferromagnetic semiconductor situation ) , and for finite band filling this corresponds to a ladder approximation in the diagram approach . similar to ( [ trans ] ) , we obtain after neglecting the magnon energies in ( [ rr ] ) the equation for the green function @xmath64 where we use the matrix notations . then we have for the operator green function @xmath65 ^{-1 } \label{gm1}\ ] ] if we consider spin dependence of electron spectrum in the simplest rigid - splitting approximation @xmath66 and thus neglect spin - dependence of the eigenfunctions @xmath67 the expressions ( [ sigmaup]),([sigmadn ] ) are drastically simplified . then the self - energy does not depend on @xmath68 and we have @xmath69 if @xmath70 is just the crystal hamiltonian ( @xmath71 , @xmath72 being the band energy ) , the expression ( [ gm1 ] ) coincides with that obtained in ref . for the hubbard model after the replacement @xmath73 . the expression ( [ gm1 ] ) can be also represented in the form @xmath74 ^{-1 } \label{gm2}\ ] ] the equation ( [ gm2 ] ) is convenient in the narrow - band case . in this limit where spin splitting is large in comparison with the bandwidth of conduction electrons we have @xmath75 and we obtain for the `` lower '' spin subband with @xmath76 @xmath77 ^{-1 } \label{nb}\ ] ] for a periodic crystal eq.([nb ] ) takes the form @xmath78 ^{-1 } \label{nb1}\ ] ] this expression yields exact result in the limit @xmath79 , @xmath80 ^{-1 } , \\ r(\epsilon)=\sum_{\bf k } \frac{1- f(t_{\mathbf k})}{\epsilon - t_{\mathbf k } } \label{nb22}\ ] ] with @xmath81 , @xmath82 the bare electron spectrum . in the limit @xmath83 eq.([nb1 ] ) gives correctly the spectrum of spin - down quasiparticles , @xmath84 ^{-1}\ ] ] with @xmath85t_{\bf k}$ ] . however , it does not describe the nqp states quite correctly , so that more accurate expressions can be obtained by using the atomic representation @xcite , @xmath86 ^{-1 } , \\ r^*(\epsilon)=\sum_{\bf k } \frac{f(t^*_{\mathbf k})}{\epsilon - t^*_{\mathbf k } } \label{nb2}\ ] ] on the other hand , the result ( [ nb ] ) , ( [ nb1 ] ) yields a good interpolation description in the hubbard model @xcite . the green s functions ( [ nb1 ] ) , ( [ nb22 ] ) , ( [ nb2 ] ) have no poles , at least for small current carrier concentration , and the whole spectral weight of minority states is provided by the branch cut ( non - quasiparticle states ) @xcite . for surface states this result was obtained in ref . in a narrow - band hubbard model . now we see that this result can be derived in an arbitrary inhomogeneous case . for a hmf with the gap in the minority spin subband nqp states occur above the fermi level , and for the gap in the majority spin subband below the fermi level . in the absence of spin dynamics ( i.e. , neglecting the magnon frequencies ) the nqp density of states has a jump at the fermi level . however , the magnon frequencies can be restored in the final result , in analogy with the case of ideal crystal , which leads to a smearing of the jump on the energy scale of a characteristic magnon energy @xmath87 . it should be mentioned once more that we restrict ourselves to the case of the usual three - dimensional magnon spectrum and do not consider the influence of surface states on the spin - wave subsystem . the expressions obtained enable us to investigate the energy dependence of the spectral density . an analysis of the electron - spin coupling yields different pictures for two possible signs of the @xmath31 exchange parameter @xmath39 . for @xmath88 the spin - down nqp scattering states form a `` tail '' of the upper spin - down band , which starts from @xmath89 ( fig.1 ) since the pauli principle prevents electron scattering into occupied states . for @xmath90 spin - up nqp states are present below the fermi level as an isolated region ( fig.2 ) : occupied states with the total spin @xmath91 are a superposition of the states @xmath92 and @xmath93 . the entanglement of the states of electron and spin subsystems which is necessary to form the nqp states is a purely quantum effect formally disappearing at @xmath94 . to understand better why the nqp states are formed only below @xmath5 in this case we can treat the limit @xmath95 t hen the current carrier is really a many - body state of the occupied site as a whole with total spin @xmath96 which propagates in the ferromagnetic medium with spin @xmath97 at any site . the fractions of the states @xmath98 and @xmath99 , respectively , so that the first number is just a spectral weight of occupied spin - up electron nqp states . at the same time , the density of empty states is measured by the number of electrons with a given spin projection which one can add to the system . it is obvious that one can not put any spin - up electrons in the spin - up site at @xmath95 therefore the density of nqp states should vanish above @xmath5 . it is worthwhile to note that in the most of known hmf the gap exists for minority - spin states @xcite . this is similar to the case @xmath88 , so that the nqp states should arise above the fermi energy . for exceptional cases with the majority - spin gap such as a double perovskite sr@xmath100femoo@xmath101 @xcite one should expect formation of the nqp states below the fermi energy . it has been proven in the previous section that the presence of space inhomogeneity ( e.g. , surface ) does not change qualitatively the spectral density picture , except smooth matrix elements . therefore further in this section we will consider , for simplicity , the case of clean infinite crystal ; all the temperature and energy dependences of the spectral density will be basically the same for the surface layer . the second term in the right - hand side of eq . ( [ n(e ) ] ) describes the renormalization of quasiparticle energies . the third term , which arises from the branch cut of the self - energy @xmath102 , describes the incoherent ( non - quasiparticle ) contribution owing to scattering by magnons . one can see that this does not vanish in the energy region , corresponding to the `` alien '' spin subband with the opposite projection @xmath53 . further on we perform for definiteness concrete calculations in the case @xmath88 ( the case @xmath90 differs , roughly speaking , by a particle - hole transformation ) . summing up eq.([n(e ) ] ) to find the total dos @xmath103 and neglecting the quasiparticle shift we obtain @xmath104 \delta ( e - t_{\mathbf{k}\uparrow } ) \nonumber \\ n_{\downarrow } ( e ) & = & 2i^{2}s\sum_{\mathbf{kq}}\frac{1+n_{\mathbf{q}}-n_{% \mathbf{k\uparrow } } } { ( t_{\mathbf{k+q\downarrow } } -t_{\mathbf{k\uparrow } % } -\omega _ { \mathbf{q}})^{2}}\delta ( e - t_{\mathbf{k}\uparrow } -\omega _ { % \mathbf{q } } ) \label{dos1}\end{aligned}\ ] ] where we consider for simplicity only second - order perturbation expression . basing on a general consideration in the previous section one can prove that , actually , this expression holds for arbitrary @xmath105 at least , in the framework of @xmath106 expansion . the @xmath107-dependence of the magnon contribution to the residue of the green s function , i.e. of the effective electron mass in the lower spin subband , and an increase with temperature of the incoherent tail from the upper spin subband result in a strong temperature dependence of partial densities of states @xmath108 , the corrections being of opposite sign . at the same time , the temperature shift of the band edge for the quasiparticle states is proportional to @xmath109 rather than to magnetization @xcite . the exact solution in the atomic limit ( for one conduction electron ) , which is valid not only in spin - wave region , but for arbitrary temperatures , reads @xcite @xmath110 in this case the energy levels are not temperature dependent at all , whereas the residues are strongly temperature dependent via the magnetization . now we consider the case @xmath6 for a finite band filling . the picture of @xmath111 in hmf ( or degenerate ferromagnetic semiconductor ) demonstrates strong energy dependence near the fermi level ( figs . if we neglect magnon frequencies in the denominators of eq.([dos1 ] ) , the partial density of incoherent states should occur by a jump above or below the fermi energy @xmath89 for @xmath88 and @xmath90 respectively owing to the fermi distribution functions . an account of finite magnon frequencies @xmath112 ( @xmath113 is the spin wave stiffness constant ) leads to smearing of these singularities , @xmath114 being equal to zero . for @xmath115 we obtain @xmath116 ( @xmath117 corresponds to the spin projections @xmath118 ) . with increasing @xmath119 tends to a constant value which is of order of @xmath120 within the perturbation theory . in the strong coupling limit where @xmath121 we have from ( [ dos1 ] ) @xmath122 in fact , this expression is valid only in the framework of the @xmath106-expansion , and in the narrow - band quantum case we have to use more exact expressions ( [ nb22]),([nb2 ] ) . in numerical calculations , we follow to ref . and smear the resolvents , @xmath123 we use the semielliptic magnon dos @xmath124 which is proportional ( with the corresponding shift ) to the bare electron dos , the maximum magnon frequency being determined by the electron concentration @xmath125 @xcite . this approximation provides the correct behavior near the fermi level ( cf . ref . ) , although gives an unphysical shift of the band bottom by the maximum magnon frequency . the results of calculations are shown in figs.3 , 4 . one can see that in the model with @xmath126 ( for @xmath127 this is equivalent to the hubbard model with the replacement @xmath128 , see refs . ) the kondo " peaks @xcite modify considerably the picture . note that the function @xmath129 , which yields dos in the lowest - order approximation in the electron concentration , does not have such peaks . to investigate details of the energy dependence of @xmath130 in the broad - band case we assume the simplest isotropic approximation for the majority - spin electrons , @xmath131 provided that we use the rigid splitting approximation @xmath132 ( @xmath133 ) , the half - metallic situation ( or , more precisely , the situation of degenerate ferromagnetic semiconductor ) takes place for @xmath134 . then qualitatively the equation ( [ alpha1 ] ) works to accuracy of a prefactor . it is worthwhile to note that , formally speaking , the nqp contribution to dos occurs also for an usual " metal where @xmath135 . in the case of small @xmath136 there is a crossover energy ( or temperature ) scale @xmath137 which is the magnon energy at the boundary of stoner continuum , @xmath138 at @xmath139 the equation ( [ dos1 ] ) for the nqp contribution reads @xmath140 \theta ( e - e_{f } ) . \label{dos111}\ ] ] at @xmath141 this gives the same results as above . however , at @xmath142 @xmath143 this contribution is proportional to @xmath144 and is _ negative _ ( of course , the total dos is always positive ) . this demonstrates that one should be very careful when discussing the nqp states for the systems which are not half - metallic . the model of rigid spin splitting used above is in fact not applicable for the real hmf where the gap has a hybridization origin @xcite . the simplest model for hmf is as follows : a `` normal '' metallic spectrum for majority electrons ( [ dosbareup ] ) and the hybridization gap for minority ones , @xmath145 here we assume for simplicity that the fermi energy lies exactly in the middle of the hybridization gap ( otherwise one needs to shift @xmath146 in the last equation , @xmath147 being the middle of the gap ) . one can replace in eq.([dos1 ] ) @xmath148 by @xmath149 , @xmath150 . first , we integrate over the angle between the vectors @xmath151 and @xmath152 it is easy to calculate @xmath153 + x\right ) \label{angle}\ ] ] where angular brackets stand for the average over the angles of the vector @xmath151 , @xmath154 here we do have the crossover with the energy scale @xmath142 which can be small for small enough hybridization gap . for example , in nimnsb the conduction band width is about 5 ev and the distance from the fermi level to the nearest gap edge ( i.e. indirect energy gap which is proportional to @xmath155 ) is smaller than 0.5 ev , so that @xmath156 . for the case @xmath157 one has @xmath158 + y+y^{3/2}=\left\ { \begin{array}{cc } y^{3/2 } , & y\ll 1 \\ y , & y\gg 1 \end{array } \right . \label{ss}\ ] ] the function @xmath159 is shown in fig.5 . thus the behavior @xmath160 takes place only for very small excitation energies @xmath161 , whereas in a broad interval @xmath142 @xmath162 one has the linear dependence @xmath163 simple qualitative considerations @xcite , as well as direct green s functions calculations @xcite for magnetic semiconductors , demonstrate that spin polarization of conduction electrons in the spin - wave region is proportional to magnetization @xmath164 a weak ground - state depolarization @xmath165 occurs in the case where @xmath88 . the behavior @xmath166 is qualitatively confirmed by experimental data on field emission from ferromagnetic semiconductors @xcite and transport properties of half - metallic heusler alloys @xcite . an attempt was used @xcite to generalize the result ( [ polar1 ] ) on the hmf case ( in fact , using qualitative arguments which are valid only in the atomic limit , see eq.([atom ] ) ) . however , we will demonstrate that the situation for hmf is more complicated . in this section we focus on the magnon contribution to dos ( [ dos1 ] ) and calculate the function @xmath167 using the parabolic electron spectrum @xmath168 and averaging over the angles of the vector @xmath151 we obtain @xmath169 where @xmath170 we have used the condition @xmath171 where @xmath172 is the spin splitting@xmath173 in the ferromagnetic semiconductor we have , in agreement with the qualitative considerations presented above : @xmath174 further on we consider the spectrum model ( [ dosbareup ] ) , ( [ dosbaredn ] ) where the gap has a hybridization origin . at @xmath175 we reproduce the result ( [ phi3 ] ) which is actually universal for this temperature region . at @xmath142 @xmath176 we derive @xmath177 this result distinguishes hmf like the heusler alloys from ferromagnetic semiconductors and narrow - band saturated ferromagnets . in the narrow - band case the spin polarization follows the magnetization up to the curie temperature @xmath178 . for finite temperatures the density of nqp states at the fermi energy is proportional to @xcite @xmath179 generally , for temperatures which are comparable with the curie temperature @xmath178 there are no essential difference between half - metallic and `` ordinary '' ferromagnets since the gap is filled . the corresponding symmetry analysis is performed in ref . for a model of conduction electrons interacting with `` pseudospin '' excitations in ferroelectric semiconductors . the symmetrical ( with respect to @xmath89 ) part of @xmath111 in the gap can be attributed to smearing of electron states by electron - magnon scattering ; the asymmetrical ( kondo - like " ) one is the density of nqp states owing to the fermi distribution function . note that this filling of the gap is very important for possible applications of hmf in spintronics : they really have some advantages only provided that @xmath180 . since a single - particle stoner - like theory leads to much less restrictive ( but unfortunately completely wrong ) inequality @xmath181 the many - body treatment of the spin - polarization problem ( inclusion of collective spin - wave excitations ) is required . now we consider an application of the results obtained above to the tunneling spectroscopy problem . the formulas of sect.4 for the energy dependence of nqp contributions are , strictly speaking , derived for the usual one - electron density of states at @xmath5 , which is observed , say , in photoemission measurements . however , the factor of @xmath182 , which is present in the expression for the tunneling current ( [ tok ] ) , does not influence the temperature dependence , and therefore these results are valid for spin polarization from tunneling conductance at zero bias in stm . the only difference in the nqp contributions to @xmath183 and @xmath103 is in that after summation over the magnon wavevector @xmath184 the integration is performed over not in the the whole fermi surface , but its two points ( see eq.([rho ] ) ) . for a spherical fermi surface for majority electrons the results differ by the constant factor of the fermi surface diameter . however , the energy and temperature dependences should be the same in a more general case . consider the bias dependence of the tunneling current for zero temperature . one can see from eq.([tok ] ) that @xmath185 again , the last proportionality can be strictly justified in the case of a spherical fermi surface only , but is qualitatively valid for arbitrary electron spectrum . one should keep in mind that sometimes the surface of hmf is not half - metallic ; in particular , this is the case of a prototype hmf , nimnsb @xcite . in such a situation , the tunneling current for minority electrons is due to the surface states only . however , the nqp states can be still visible in the tunneling current via the hybridization of the bulk states with the surface one . the hybridization lead to the fano antiresonance picture which is usually observed in stm investigations of the kondo effect at metallic surfaces ( see , e.g. , refs . ) . in these cases the tunneling conductance will be proportional to a mixture of @xmath186 and @xmath187 , @xmath188 being the real part of the on - site green s function , @xmath189 ( @xmath190 stands for principal value , @xmath191 is referred to the fermi energy ) . surprisingly , in this case the effect of nqp states on the tunneling current can be even more pronounced in comparison with the ideal crystal . the reason is that the analytical continuation of the jump in @xmath103 is logarithm ; both singularities are cut at the energy @xmath192 nevertheless , the energy dependence of @xmath188 can be pronounced , see fig.6 . this is similar to the effect of enhancement of the nqp contribution to the x - ray absorption and emission spectra , which was predicted in ref . . now we discuss in more detail the energy dependence for @xmath193 . the analytical continuation of the @xmath194-contribution to @xmath103 yields the contribution @xmath195 in @xmath188 which is non - zero on the other side with respect to @xmath5 ( a situation that is formally similar to the electronic topological transition , see ref . ) . the one - sided linear dependence in @xmath103 according to eq.([ss ] ) corresponds to @xmath196 in @xmath197 . stm measurements of electron dos give also an opportunity to probe _ bosonic _ excitations interacting with the conduction electrons . due to electron - phonon coupling , the derivative @xmath198 and thus @xmath199 at @xmath200 have peaks at the energies @xmath201 corresponding to the peaks in the phonon dos . according to our results ( see , e.g. , eq.([dos1 ] ) , the same effect should be observable for the case of electron - magnon interaction . however , in the latter case these peaks are essentially asymmetric with respect to the fermi energy ( zero bias ) due to asymmetry of the non - quasiparticle contributions . this asymmetry can be used to distinguish phonon and magnon peaks . in the present paper we have demonstrated that non - quasiparticle states in half - metallic ferromangnets exist not only for an ideal crystal , but also in the presence of an arbitrary external potential . in particular , they occur at the surface of the half - metallic ferromagnets . these states can be probed by the stm both directly and via their effect on the surface states ( the fano antiresonance case ) . therefore , they can be observable even for the situation of surface `` dead layers '' where the surface is not half - metallic . the expressions obtained can be used for realistic electronic structure calculations of nqp contributions to the electron energy spectrum of the surfaces of hmf . temperature dependence of the spin polarization at the fermi energy which can be also probed by the stm follows the temperature dependence of magnetization an very low temperatures . for the hmf with a hybridization gap , there is a crossover energy ( temperature ) @xmath202 where the character of the temperature dependence is changed . the energy dependence of the nqp contributions ( and consequently the bias dependence of the tunneling current ) is strongly influenced by the band structure too . in particular , for hmf with a hybridization gap this demonstrates a linear rather than @xmath203 behavior in a wide interval . in the narrow - band case a kondo - like peak ( fig.4 ) near the fermi level should be observed in tunneling experiments . due to asymmetry of nqp states with respect to the fermi energy , the magnon peaks in @xmath204 are also asymmetric with respect to the zero bias , in contrast with the phonon ones . this gives an opportunity to distinguish between phonon and magnon peaks in the inelastic spectroscopy by stm . the research described was supported in part by grant no . 747.2003.2 ( support of scientific schools ) from the russian basic research foundation and by the netherlands organization for scientific research ( grant nwo 047.016.005 ) . r. wiesendanger , h .- j . guentherodt , g. guentherodt , r.j . cambino , and r. ruf , phys . lett . * 65 * , 247 ( 1990 ) ; s. heinze , m. bode , a. kubetzka , o. pietzsch , x. nie , s. blgel , and r. wiesendanger , science * 288 * , 1805 ( 2000 ) ; m. kleiber , m. bode , r. ravli , and r. wiesendanger , phys . lett . * 85 * , 4606 ( 2000 ) . exchange model of a half - metallic ferromagnet with @xmath205 for the semielliptic bare band with the width of @xmath206 . the fermi energy calculated from the band bottom is 0.15 ( the energy is referred to @xmath5 ) . the magnon band is also assumed semielliptic with the width of @xmath207 . the non - quasiparticle tail of the spin - down subband ( lower half of the figure ) occurs above the fermi level . the corresponding picture for the empty conduction band is shown by dashed line ; the short - dashed line corresponds to the mean - field approximation . ] , other parameters being the same as in fig.1 . the spin - down subband ( lower half of the figure ) nearly coincides with the bare band shifted by @xmath208 . non - quasiparticle states in the spin - up subbands ( upper half of the figure ) occur below the fermi level ; the short - dashed line corresponds to the mean - field approximation . ]
the role of the many - body ( spin - polaronic ) effects in the scanning tunneling spectroscopy of half - metallic ferromagnets ( hmf ) is considered . it is shown that the non - quasiparticle ( nqp ) states exist in the majority or minority spin gap in the presence of arbitrary external potential and , in particular , at the surfaces and interfaces . energy dependence of the nqp density of states is obtained in various models of hmf , an important role of the hybridization nature of the energy gap being demonstrated . the corresponding temperature dependence of spin polarization is calculated . it is shown that the nqp states result in a sharp bias dependence of the tunneling conductance near zero bias . asymmetry of the nqp states with respect to the fermi energy provides an opportunity to separate phonon and magnon peaks in the inelastic spectroscopy by stm .
the knowledge of the nature of the matter perturbations in the observed universe is crucial for obtaining information about the very early universe and the very high energy physics . the deepest presently available window to the early universe is the cosmic microwave background ( cmb ) , that in the next decade will be deeply investigated by the map and planck missions , after the discovery of its anisotropies by cobe @xcite . however , the cosmic gravitational background ( cgb ) will play a crucial role in the next future as the most powerful tool in reconstructing primordial physics ( see @xcite for an extensive overview ) . infact , it consists of the gravitational waves ( gw ) generated during the inflationary era ; they are carriers of unperturbed physical traces of the very primordial history of the universe , since the gw decoupling probably occurred about 70 @xmath1foldings before the recombination ! the study of the perturbations and defects produced during inflation underwent recently a great revival ; it is motivated by the evidence of strong inhomogeneities from the direct reconstruction of the three dimensional matter distribution traced by galaxies and their peculiar velocities in the modern redshift surveys . the most important conclusions of these observations are that strong underdensities , or voids , appear very prominent in the data @xcite@xcite@xcite@xcite . in search for the inflationary generation mechanisms for such inhomogeneities , one of the most interesting ideas introduced in cosmology in recent years is the possibility of performing a phase transition _ during _ inflation . in such scenarios , two fields act on stage : one , say @xmath3 , slow rolls , driving enough inflation to solve the standard problems ; the second field , say @xmath4 , tunnels from a false vacuum state to an energetically favoured true vacuum state , producing bubbles of the new phase embedded in the old one . both processes are governed by a two - field potential @xmath5 . to avoid the graceful exit problem , the true vacuum state has to allow for a period of inflation on its own . we can then speak of a true vacuum channel over which the bubbles slow roll until inflation ends , and reheating takes place . depending on the potential , we can distinguish two different scenarios of first order inflation . the first is the classical extended inflation @xcite : the bubbles are produced in a copious quantity , so that they eventually fill the space and complete the transition . to avoid too large distortions on the cmb , this scenario must produce very small bubbles @xcite , so that they are rapidly thermalized after inflation . no trace of the bubbles is left in our universe , and from this point of view such scenarios do not lead to new predictions over inflation without bubble production . in the second scenario , proposed in @xcite and implemented in @xcite , the phase transition is completed _ before _ the end of inflation ; the amount of @xmath6-foldings between the phase transition and the end of inflation makes the scale of the bubbly perturbations interestingly non - vanishing , and allows them to leave observable traces . in particular , if a phase transition occurred sufficientely early , the bubbles are streched to cosmological scales , and the present large scale structure is therefore strictly linked to the primordial originating transition , which is therefore observable and testable as firstly suggested in @xcite . in this work we concentrate upon the traces of the above phaenomenology on the cgb . the most natural way for the nucleated bubbles to generate gravitational waves is through collision with each other ; compared with the ordinary tensor perturbations occurring in slow rolling inflation , this is infact a very potent source of primordial gw , as firstly argued in @xcite . the problem was analysed in the context of the ordinary extended inflation , in which the bubbles are completely empty and the phase transition occurs at the end of inflation . the computation of the amplitude and frequency spectrum of the gravitational radiation was performed firstly in the case of two bubbles @xcite and then by considering an envelope of hundreds of bubbles @xcite , that substantially confirmed the previous results ; a work emphasising the observation possibilities can be found in @xcite . finally in @xcite the problem of gravitational waves from lower energy first order phase transition ( like the electro - weak transition occurring nearly at 100 gev ) was considered . here we want to extend these results to the second kind of inflationary phase transitions mentioned above ; the main differences from the treated cases are that @xmath7 ) the nucleation epoch occurs before the end of inflation , and @xmath8 ) generally the bubble are not empty . the paper is organized as follows : in section ii we recall the main results on the gravitational radiation emitted during the first order phase transition in the scenario of extended inflation , pointing out the approximations for which the computations and the results are valid ; in section iii we extend these results to recently proposed inflationary models capable of performing a first order phase transition before the end of inflation and we discuss the observation possibilities and the existing constraints ; finally , section iv contains the conclusions . a detailed analysis of a first order phase transition in the context of extended inflation can be found in @xcite . the field @xmath4 that undergoes the transition is the same that drives inflation , and its dynamics is assumed to be governed by a potential containing two non - degenerate minima , the true and false vacuum ( tv , fv ) . the central quantity needed to characterize the transition is the bubble nucleation rate for unit volume in the semiclassical limit @xcite @xmath9 where @xmath10 is the euclidean least action over the bounce minus the action for the external desitter spacetime solution @xcite . the constant @xmath11 ( with a dimension of mass ) is of the order of the energy @xmath12 at which the phase transition occurs . values of @xmath13 a few orders of magnitude below the planck mass @xmath14 are generally assumed , in order to avoid quantum gravity effects @xcite . the transition is completed roughly when at least one bubble per unit hubble volume is nucleated , and accurate computations @xcite show that extended inflation is successful ( in the sense that bubbles percolate ) if at a time @xmath15 @xmath16 the bubbles are empty and their walls rapidly approach the light speed . depending on the intensity of the nucleation rate , the collisions between the bubbles occur when they have comoving spatial dimension less or equal to the effective horizon @xmath17 at the transition epoch given by @xmath18 . in the context of the standard extended inflation almost all the bubbles are nucleated at the end of inflation , so that if we take @xmath19km / sec / mpc in an @xmath20 universe their comoving size is approximatively @xmath21 . such bubbles soon reenter the horizon and rapidly thermalize without leaving trace in the matter distribution . on the other hand , a very characteristic gw spectrum is produced during the transition . if the bubbles are perfectely spherical , no spacetime perturbation is seen from the outside . when two ( or more ) bubbles collide they become a gw source , and the collection of collisions during the whole phase transition leaves observationally interesting traces on the cgb @xcite . the problem of computing the emitted gw spectrum is conceptually simple but computatively difficult . one has to simulate the bubble nucleations with rate given by ( [ tunnrate ] ) and to evolve them with the klein - gordon equation . whenever collisions occur , one must compute the radiation emitted . it has been performed @xcite with the aid of two main approximations : linearized gravity and static background . the former is consistent if the fraction of energy that goes in gw is small with respect to unity ; the latter is valid if the transition completes within an hubble time . following @xcite , the total energy radiated at frequency @xmath3 in the direction @xmath22 into the solid angle @xmath23 and in the interval @xmath24 is @xmath25 where @xmath26 is the projection tensor for gravitational radiation @xcite and @xmath27 the fourier transform of the energy momentum tensor . the computations were firstly performed in the simplified case of two colliding bubbles @xcite . the end of the transition in that case was modelled as a modulating function @xmath28 that makes the signal vanish after a cutoff time @xmath29 of the order of the initial separation between the bubbles . in @xcite an ingenious model for @xmath30 was introduced by excluding from the integration the spatial region where bubbles overlap ; at @xmath31 the bubbles completely fill the space , and the gw emission ends . this also allowed to perform many bubbles simulations . the results are intuitive and consistent with the approximations . firstly , the bounce is approximated as @xcite @xmath32 where @xmath33 sets the natural scale for the phase transition . since this is of the order of the initial average separation between the colliding bubbles , the static backgorund approximation requires that @xmath34 . the gw spectrum shows a very characteristical peak at @xmath35 and the fraction of energy radiated in gw is found @xmath36 ; @xmath37 is the critical energy density at the time @xmath15 ; @xmath38 counts the number of relativistic degrees of freedom at the temperature @xmath39 of the phase transition , and is typically of the order of 100 . we remark that these results agree with physical expectations : the peak frequancy is simply the inverse of the time scale of the process , and the @xmath40 scaling for @xmath41 can be naturally inferred in terms of the energy density @xmath42 and @xmath33 @xcite . to obtain the corresponding present quantities , one has to take into account the cosmic redshift from @xmath39 down to the present 3 k. @xmath41 remains unchanged during the radiation epoch , but undergoes a decrease during the matter dominated era ; @xmath43 redshifts via the cosmic expansion . the quantities relevant at the present are therefore @xmath44 @xmath45 where we have defined @xmath46 and @xmath47 . we emphasize once again that the above results hold for the extended inflation scenario , in which bubbles are nucleated at the end of inflation and rapidly thermalize . in the next section we will extend these results to the models of first order inflation capable to perform the nucleation epoch well before the end of inflation , showing how the above results change and discussing the consequences on the observability . recent studies on inflation @xcite have shown that a not @xmath48 slow rolling is generated by the gravitational part of the lagrangian , already at the level of fourth order gravity ( fog ) . a matter field potential containing false and true vacuum minima generates a first order phase transition , but the latter in general occurs at an epoch characterized by a number @xmath49 of @xmath1foldings before the end of inflation . depending on this quantity , the scale of the bubble - like perturbations may be stretched out to cosmological size by the exponential growth , thus candidating the bubbles to be the seeds for the formation of the voids observed today @xcite . we consider the consequences on the cgb spectrum of this fog model ( and we refer to @xcite and @xcite for an extensive treatment ) ; since a conformal transformation makes the lagrangian as in ordinary two field inflation , our results hold for any kind of inflation in which one field slow rolls and the other one undergoes a first order phase transition . once the conformal transformation is performed , the fog action takes the form @xmath50\,,\ ] ] where the potential is @xmath51\,,\ ] ] and generates tv bubbles with @xmath52 , a degenerate quartic , and @xmath53 , the symmetry breaking term . the slow roll inflation driven by @xmath3 takes place at @xmath54 , and is over when @xmath3 approaches zero . at large @xmath3 , the potential @xmath55 is dominated by @xmath56 , and thus the false vacuum minimum at @xmath57 , for which @xmath58 $ ] is unstable with respect to tunneling towards the true vacuum @xmath59 ( for which @xmath60 ) . at small @xmath3 , instead , @xmath55 is dominated by @xmath61 , and both the true and the false vacua converge to the global zero - energy minimum at @xmath62 , where inflation ends and reheating takes place . the slow - roll solution in this model for @xmath54 can be written very conveniently as @xcite @xmath63 where @xmath64 is the number of @xmath6-foldings to the end of inflation . it has been shown @xcite that taking into account the gravitational corrections and dropping the thin wall approximation due to a finite thickness @xmath65 for the wall surrounding a bubble of radius @xmath66 , the nucleation rate @xmath67 defined in ( [ tunnrate ] ) becomes @xmath68 and in terms of @xmath64 the bounce is @xmath69\left[1-\left({n\over n_3 } \right)^4\right]\ ] ] where the first brackets is the thin wall correction , and the second brackets is the gravitational correction ; moreover , the @xmath70 are related to the physical quantities by @xmath71 by indicating with @xmath49 the amount of @xmath1foldings to the end of inflation when the phase transition takes place , the quantity @xmath72 defined in ( [ condition ] ) takes the approximate form @xmath73\left[1-\left({n_0\over n_3 } \right)^4\right]\right\}\ \ .\ ] ] the two approximations that have been adopted to obtain the above formulas are satisfied by the conditions @xmath74 ( thin - wall ) , and for @xmath75 ( gravitational correction ) . also , @xmath76 is required to guarantee tunneling through tv bubbles @xcite . the central underdensity of a bubble is determined by the shape of the potential ( [ conpot ] ) at the nucleation epoch , characterizes by @xmath64 : @xmath77^{-1}\ \ , \ \ n_4 ^ 2=3\pi { \psi_0 ^ 2 m^2\over m^2}={64\over 9\pi } m^{2}n_{1}^{-4}n_{3}^{8}\ \ .\ ] ] from the current microwave background measurements , we obtain @xmath78 ( in planck units ) @xcite , a value that we will adopt in the results below . however , since in our model one should also consider the contribution of the bubbles to the microwave background , this constraint is actually only an upper limit on @xmath79 . as an intuitive feature , note that the four physical parameters @xmath80 , and therefore the four @xmath6-folding constants @xmath81 fully determine the inflationary potential at @xmath49 ( the nucleation epoch ) : the vacuum energy density , the energy difference between fv and tv , the amplitude of the barrier and the value @xmath82 of the fv phase . the linearized gravity approximation involved to obtain the results ( [ omeganow],[omegagwnow ] ) is surely satisfied in our case since the energy that goes into the walls and therefore in gw is a fraction @xmath83 given by ( [ delta ] ) of the total fv energy in the bubble . the other important approximation is to consider spacetime as static during the transition ; in other words , the time scale of the nucleation era must be smaller than the hubble time . this is satisfied in our case if the condition @xmath34 holds at @xmath49 , where @xmath84 is defined in ( [ beta ] ) ; this means @xmath85 , that using ( [ bn ] ) becomes @xmath86\ge 1\ \ .\ ] ] by defining @xmath87 the quantity in brackets in ( [ fasttrans ] ) is simply @xmath88 still taking into account the gravitational and post - thin wall corrections . the range of interest for both @xmath89 and @xmath90 is @xmath91 $ ] for consistency of our approximations . since @xmath92 does not appeare in @xmath93 , for the validity of ( [ fasttrans ] ) it is enough that @xmath93 assumes some positive value ; it is easily seen that this is true in the range @xmath94 . by choosing a value of @xmath49 and @xmath95 to fix @xmath89 , a value of @xmath90 always exists that yields @xmath96 ; therefore , our model performs a fast transition for @xmath97^{1/4}\ \ .\ ] ] we can now extend the results ( [ omeganow],[omegagwnow ] ) to the present inflationary models . the peak frequency of the gw spectrum contains an @xmath98 redshift due to the inflationary expansion between the phase transition and the end of slow rolling . moreover , in calculating @xmath41 we must take into account that the fraction of energy that goes in gravitational radiation is reduced by a fraction @xmath83 , ( [ delta ] ) , with respect to ( [ omegagwnow ] ) . therefore , the results are @xmath99 @xmath100 in the following we will adopt the values @xmath101 ( fixing the epoch of the end of slow rolling at typical gut energy scales ) , and @xmath102 . from the above formulas it is clear that in general any model of two field first order inflation radically changes the frequencies of the gw spectrum to be observed today with respect to the ordinary extended inflation models . the main difference is represented by the shift of the peak frequency towards lower values depending on the amount of @xmath1foldings between the phase transition and the end of inflation . this has very interesting consequences for what concerns the search for the present traces of the era driven by very high energy physics . we make now some general considerations about this , referring to the next two sub - sections to a detailed analysis of two important cases . formulas ( [ gwfog1],[gwfog2 ] ) may be rewritten emphasizing the relation between the physical parameters @xmath103^{2 } } \exp ( -2n_{0})\ \ , \ ] ] with the constraint @xmath104 . in fig.1 we plot experimentally interesting @xmath105 as a function of the peak frequency @xmath43 and for different @xmath49 ( the change in grey tonality corresponds to a scansion of 5 in @xmath49 , as indicated on the legend box ) ; for clearness , we have set @xmath106 ( top ) and .01 ( bottom ) . each point on the figure correspond to the gw spectrum generated by a first order phase transition occurred @xmath49 @xmath6-foldings before the end of inflation . for @xmath107 we obtain the known results in the case of extended inflation ( [ omeganow],[omegagwnow ] ) . the increase of @xmath49 pushes the peak frequency towards exponentially small values and intersects very interestingly the plotted expected level of sensitivities of the next generation of gw interferometric detectors : if first order phase transitions occurred at @xmath108 during the inflationary era ( realistically arising from spontaneous breaking of very high energy symmetries ) , they should leave next future detectable peaks in the cgb frequency spectrum . by continuing to increase @xmath49 beyond several decades , we enter in the class of phase transitions that could leave traces in the present large scale matter distribution . in particular , at @xmath109 , the remnants of the nucleated bubbles are of astrophysically interesting size , and correspond to the large voids detected in the galaxy distribution @xcite @xcite . for such models , the gw frequencies correspond to cosmological wavelenghts , to be interestingly investigated through their induced cmb anisotropies ; as we will show in the sequel , a deep insight already comes from considering the anisotropy amplitude ; however it is interesting to point out that in the future these models could be tested with powerful methods ( presently under investigation ) capable to extract the pure cgb signal from the data coming from the next high resolution cmb experiments @xcite . in the following we consider in more detail the correspondence between the cgb and the parameters of the effective lagrangian , focusing on two important cases : spectra detectable by the next ligo interferometric detectors , and spectra from colliding bubbles of astrophysically interesting size . the advanced ligo project will reach a sensibility of about @xmath110 in the frequency range @xmath111 hz ( see @xcite and references therein ) . from ( [ gwfog1],[gwfog2 ] ) , we see that the amplitude and the peak frequency fix the products @xmath112 and @xmath113 , namely two conditions for our physical parameters . then , by fixing @xmath114 so to yield a reasonable parametric region for @xmath115 ( we remember that @xmath116 ) , we may search for the physical parameter set that could explain the observation of a definite peak at some frequency in the above ranges . such set is constrained by ( [ gwfog1 ] ) , @xmath117^{1\over 4}\le n_{1}\le \left[{3\cdot 10^{7}\over \omega_{min}}4n_{0}^{3 } z(x , y)e^{-n_{0}}\right]^{1\over 4}\ \ , \ ] ] to have the peak frequency in the observation range , and by ( [ gwfog2 ] ) , @xmath118^{1\over 4}\ \ , \ ] ] to have a sufficient amplitude . moreover ( [ fasttrans2 ] ) ( with @xmath96 of course ) must be satisfied to have the completion of the transition in less than an hubble time and @xmath119 to have nucleation of tv bubbles . the dashed areas in fig.2 are the regions of the parameters @xmath120 for which both the peak frequency and the amplitude fall into the observation ranges ; the upper panel shows the limit in which the post - thin wall and gravitational corrections are negligible , while in the lower panel such corrections are important . the constraints define the region of observability shaded in fig.2 ; as an interesting and intuitive feature we note that a first order inflationary phase transition , that is ligo detectable , occurs at @xmath121 @xmath1foldings before the end of inflation , that is just the amount necessary to redshift the corrisponding frequencies in ordinary extended inflation ( @xmath122 , see ( [ omeganow ] ) and fig.1 ) towards the observation range . from to the nucleation epoch it is easily found that the comoving size of these perturbations is approximatively @xmath123 ; this means that they reenter the horizon early in the radiation dominated era and most likely do not leave traces in the density field . finally we remark once again that for each set of @xmath81 there exists a set of physical parameters @xmath124 , as it is easily seen from ( [ tunnrate2],[bn ] , [ n],[q ] ) . as we already mentioned in the introduction , the recent analysis on the modern redshift surveys have assessed the large scale voids as the dominant feature of the nearby universe . several works have been performed in these years on the hypothesis that the primordial origin of the voids is in an inflationary phase transition in the context of models of fog and in general two field inflation @xcite . here we add a new tool . it has been shown @xcite that voids of tens of mpc of diameter would be the relics of bubbles with comoving size expanded by a factor @xmath125 between the phase transition and the end of inflation . this request fixes the product @xmath126 and by requiring @xmath127 we have @xmath109 . the frequency in ( [ gwfog1 ] ) is therefore : @xmath128 on these frequencies there is the most stringent presently available constraint upon the amplitude of the cgb , provided by the amount of anisotropy of the cmb ( see @xcite and references therein ) . precisely , this constraint imposes that @xmath129 and consequentely our frequencies localize themselves on the high frequency border of this range , where @xmath130 ( @xmath19 km / sec / mpc ) . note that the corrisponding wavelenghts determine cmb anisotropies on small angular scales ( @xmath131 ) , that will be deeply investigated by the map and planck missions of the next decade , as mentioned in the introduction . a great effort is being done in order to develope theoretical analysis instruments capable to extract the pure cgb signal from the whole spectrum of cmb anisotropies , both for what concerns pure @xmath132 and polarization ( see e.g. @xcite ) ; this matter will provide a powerful tool to investigate the whole cgb spectrum , and in particular the traces of early inflationary phase transitions , that are the subject here . here we concentrate on the overall amplitude constraint given by ( [ allen ] ) , since it already provides a deep insight into the physical parameter space ; from ( [ delta],[gwfog2 ] ) we see that it must be @xmath133 or , alternatively to ( [ l0 ] ) , @xmath134^{1\over 4}\ \ ; \ ] ] this sets a very narrow range ( at fixed @xmath135 and @xmath136 ) for @xmath49 and @xmath92 . moreover , another condition arises from ( [ delta ] ) for the minimal @xmath137 capable to yield the voids observed today @xcite @xmath138 in fig.3 we plot again all the constraints for @xmath139 ( upper panel , negligible post - thin wall and gravitational corrections ) and @xmath140 ( lower panel ) . in both panels , each curve refers to the indicated constraint : the solid line represents the condition ( [ g ] ) , and guarantees that the bubbles are expanded to the observed sizes . the short dashed line draws the condition ( [ void4 ] ) . finally , the long dashed line represents the strong constraint coming from the cmb ( [ void2],[void3 ] ) . the part of solid line below all the curves satisfies all the constraints , and restricts to @xmath141 the epoch of the phase transition ; also , the gravitational and post - thin wall corrections ( important in the lower panel ) allow for a viable parameter set slightly larger than when they are negligible ( upper panel ) . for physical parameters on the allowed set , the presently large scale observed voids correspond , both for size and underdensity , to the tv bubbles nucleated during inflation , without exceeding the constraints coming from the cmb isotropy . we have extended the known results on the stochastic gravitational background produced by colliding bubbles in two fields models of first order inflation . in such models , a field performs the first order phase transition and a second field ( that in fourth order gravity is of gravitational origin ) provides the inflationary slow rolling . the resulting general phaenomenology is that the phase transition occurs well before the end of inflation . this has very non - trivial consequences on the gravitational radiation produced by colliding bubbles . it is found that the expansion between the phase transition and the end of inflation cause an additional redshift for the peak frequency of the spectrum with respect to the ordinary models of extended inflation ; also , the energy carried by the gravitational waves is reduced by a fraction equal to the density contrast of the ( not completely empty ) bubbles . the result is that if first order phase transitions occurred during the inflationary era ( realistically arising from spontaneous breaking of very high energy symmetries ) , they should leave next future detectable peaks in the cgb frequency spectrum . that detection would be a first experimental verification of our very high energy physics theoretical investigations . in particular , a peak at some frequency explored by the next ligo interferometers may be explained in terms of a first order phase transition occurred typically 10@xmath020 @xmath1foldings before the reheating at gut energy scales ( @xmath142 gev ) . moreover , the gravitational radiation produced in recently proposed models capable of generating the observed large scale voids , has been examined . in these models , the phase transition occurs more than 50 @xmath1foldings before the end of inflation . as expected , the typical frequencies of the spectrum are in the range in which the isotropy of the cmb puts the strongest existing constraint on the amplitude of the gravitational background , and gives a very deep insight into the physical parameter space . we find a precise new relation among the physical parameters in the class of the models in which the phase transition is fast ( namely less than an hubble time ) ; the cmb constraint requires that the phase transition occurs generally more than 55 @xmath1foldings before the end of inflation . these results localize the parameter set of the effective lagrangian for which the present large scale observed voids correspond , both in size and underdensity , to the tv bubbles nucleated during the inflationary era , without perturbing the cmb . the fraction of cmb anisotropies and polarization due to phaenomena investigated here could be recognized in the high resolution cmb data that will be provided by the experiments of the next decade . b. allen 1996 _ proceedings of the les houces school on astrophysical sources of gravitational waves _ , cambridge university press ; gr - qc/9604033 c.l . bennet , a. kogut , g. hinshaw , a.j . banday , e.l . wright , k.m . gorski , d.t . wilkinson , r. weiss , g.f . smoot , s.s . meyer , j.c . mather , p. lubin , k. loewenstein , c. lineweaver , p. keegstra , e. kaita , p.d . jackson & e.s . cheng 1994 _ ap.j . _ * 436 * , 423 u. seljak and m. zaldarriaga , 1997 to appear in _ phys . _ ; astro - ph/9609169 m. kamionkowski and a. kosowsky , 1997 submitted to _ phys . d _ ; astro - ph/9705219 vogeley m. s. , geller m. j. & hucra j. p. 1991 _ ap.j . _ * 382 * , 54 ; kauffmann g. & fairall a. p. 1991 _ mon.not.r.astr.soc . _ * 248 * , 313 el - ad h. , piran t. & da costa l. n. , 1996 _ ap.j . _ * 462 * , l13 el - ad h. , piran t. & da costa l. n. , 1996 submitted to , astro - ph/9608022 da costa l.n . , freudling w. , wegner g. , giovanelli r. , haynes m. , and salzer j.j . 1996 _ ap.j . _ * 468 * , l5 f. occhionero , c. baccigalupi , l. amendola & s. monastra 1997 submitted to _ phys . d _ c. baccigalupi , l. amendola & f. occhionero , 1996 , preprint astro - ph/9603064 , _ mnras _ in press c. baccigalupi , l. amendola & f. occhionero 1996 , work in progress d. la and p.j . steinhardt 1989 _ phys . lett . _ * 62 * , 376 d. la , p.j . steinhardt and e. bertschinger 1989 _ phys . b _ * 231 * , 231 ; e.j . weinberg 1989 _ phys . d _ * 40 * , 3950 ; t. damour , g.w . gibbons and c. gundlach 1990 _ phys . _ * 64 * , 123 ; p.j . steinhardt and f.s . accetta 1990 _ phys . _ , * 64 * , 2740 ; r. holman _ _ 1990 _ phys . b _ _ 237 _ , 37 f.c . adams and k.freese 1991 _ phys d _ * 43 * , 353 m.s . turner and f. wilczek 1990 _ phys . _ * 65 * , 3080 m.s . turner , e.j . weinberg and l.m . widrow 1992 _ phys . * 46 * , 2384 a.r . liddle and d. wands 1991 _ mnras _ * 253 * , 637 f. occhionero & l. amendola 1994 _ phys . d _ * 50 * , 4846 d. la 1991 _ phys . b _ * 265 * , 232 l. amendola and f. occhionero 1993 _ ap.j . _ * 413 * 39 l. amendola , c. baccigalupi , r. konoplick , f. occhionero & s. rubin 1996 _ phys . * 54 * 7199 s.weinberg , _ gravitation and cosmology _ , wiley , new york , 1972 a. kosowsky , m.s . turner & l.m . widrow 1992 _ phys . d _ * 45 * , 4514 a. kosowsky & m.s . turner 1993 _ phys . d _ * 47 * , 4372 a. kosowsky , m.s . turner & l.m . widrow 1992 _ phys . lett . _ * 69 * , 2026 m. kamionkowski , a. kosowsky & m.s . turner 1994 _ phys . d _ * 49 * , 2837 s.coleman 1977 _ phys . d _ * 15 * , 2929 ; c. callan & s. coleman 1977 _ ibid . _ * 16 * 1762 ; s. coleman & f. de luccia 1980 _ ibid . _ * 21 * , 3305 d.s . goldwirth & h.w . zaglauer 1991 _ phys . lett . _ * 67 * , 3639 * fig.1 . amplitude - peak frequency relation for the cgb generated in a first order phase transition occurred @xmath49 @xmath1foldings before the end of inflation . for clearness , we have exploited the two interesting cases @xmath143.1 ( top ) and .01 ( bottom ) and we have evidentiated the dependence of the curves on @xmath49 ( the grey tonality change correspond to a scansion of 5 in @xmath49 , as indicated in the legend box ) . note the shift of the frequencies towards ranges that will be investigated by the next interferometric detectors . . advanced ligo detection region for ( @xmath144 ) , in the cases for which the post - thin wall and gravitatinal corrections are negligible ( upper panel ) or not ( lower panel ) . the dashed areas contain physical parameters respecting all the indicated constraints ; particularly it yelds frequencies among the maximal and minimal limits ( long broken lines ) and amplitude above the minimal observable one ( solid lines ) . * insight upon the physical parameters of effective lagrangian yielding large scale structure from bubbles ; the cases for which the post - thin wall and gravitational corrections are negligible ( upper panel ) or not ( lower panel ) are shown . the segments of the solid line below all the other curves sketch the relation among the physical parameters for models capable of explaining the large scale observed voids as the relics of bubbles nucleated more than 55 @xmath1foldings before the end of inflation ; all the performed constraints are respected , and the @xmath92 axis is logaritmic to show all them ; note particularly the request of sufficientely deep bubbles ( short broken lines ) and the strong constraint coming from the upper limit to the level of cmb anisoptropy ( long dashed lines ) .
we consider true vacuum bubbles generated in a first order phase transition occurring during the slow rolling era of a two field inflation : it is known that gravitational waves are produced by the collision of such bubbles . we find that the epoch of the phase transition strongly affects the characteristic peak frequency of the gravitational waves , causing an observationally interesting redshift in addition to the post - inflationary expansion . in particular it is found that a phase transition occurring typically 10@xmath020 @xmath1foldings before the reheating at @xmath2 gev may be detected by the next ligo gravity waves interferometers . moreover , for recently proposed models capable of generating the observed large scale voids as remnants of the primordial bubbles ( for which the characteristic wave lengths are several tens of mpc ) , it is found that the level of anisotropy of the cosmic microwave background provides a deep insight upon the physical parameters of the effective lagrangian . 9.1 in 6.5 in -.45 in
in recent years , a series of young stellar kinematic groups ( clusters , associations , and moving groups ) of late - type stars with similar space motion and ages ranging from 8 to 50 myr ( see * ? ? ? * and references therein ) has been discovered in our neighbourhood : tw hya , @xmath0 pic , ab dor , @xmath1 cha , @xmath2 cha , tucana and horologium associations . in addition , several more distant young associations such as mbm 12 @xcite , corona australis @xcite , and possibly the group of stars with a motion similar to that of hd 141569 @xcite have also been identified . in the galactic velocity space , they situate inside the boundaries of the ( see fig . [ fig : uv ] ) , a mixture of young stellar complexes ob and t - associations and clusters with different ages @xcite . these associations of very young stars are excellent laboratories for investigations of forming planetary systems @xcite . nevertheless , they are generally situated at distances above 50 pc , which makes them less accessible to adaptive optics systems even on large telescopes . it is well - known that tightly bound , long - lived open clusters can account for only a few per cent of the total galactic star formation rate @xcite . therefore , either most clusters and associations disperse very quickly after star formation has started or most are born in isolation @xcite . the existence of very young moving groups ( mgs ) with a few dozens of stars showing the same spectroscopic properties i.e. age , metallicity , level of magnetic activity is in agreement with the first explanation . small associations of stars may be dispersed by galactic differential rotation since they are not gravitationally bounded enough , taking into account that their nucleus consist of only a few stars as in the case of the ursa major mg ( see king et al . 2003 , for a recent review ) or the recently discovered ab doradus mg @xcite . the location of these young mgs inside the local association and its proximity in the @xmath3-plane can be explained as the result of the juxtaposition of several star forming bursts in adjacent cells of the velocity field ( see montes et al . 2001 , and references therein ) or dynamical perturbations caused by spiral waves @xcite . thus , one expects to find groups of coeval stars with similar space motion in our neighbourhood . in 2001 , the @xmath0 pic mg @xcite a group of stars with an age of @xmath4 12 myr @xcite at a mean distance of @xmath4 35 pc co - moving with the well - known young star @xmath0 pic was confirmed to be the closest kinematic group up to date . more recently , @xcite have identified a new group of stars co - moving with the also well - known young star ab dor , at a mean distance of @xmath4 30 pc , and with an age of @xmath4 50 myr . nevertheless , the existence of a nearer association of a few stars was proposed by @xcite and studied in detail by @xcite , though its existence is quite controversial . here we discuss about the fact of the nearest mgs using both spectroscopic and photometric criteria of membership for a sample of stars that includes the proposed members from the literature and our list of young cool stars possible members of young stellar kinematic groups @xcite . based on the kinematics of young solar analogues in the solar neighbourhood , @xcite confirmed the existence of a group of four stars ( marked with @xmath5 in table [ tab : mgs ] ) co - moving in the space towards the constellation of hercules . recently , @xcite has extended the sample of late - type stars of this mg up to 15 nearby ( d @xmath6 25 pc ) candidates , proposing the name hercules - lyra since several members show a radiant `` _ evenly matched _ '' with this constellation . comparing the level of chromospheric activity of the stars of his sample with that of the members of the ursa major association and looking for the existence of lithium in their spectrum , he notices that several candidates of hercules - lyra appear to be coeval of the ursa major stars , for which he gives an age of @xmath4 200 myr . on the contrary , other candidates seem to be older ( e.g. hd 111395 ) or younger ( hd 17925 , hd 82443 , and hd 113449 ) , questioning the existence of hercules - lyra as an entity independent of the local association . however , he considers unlikely that the majority of his sample can originate from the pleiades alone , or other clusters of the local association since `` _ they are poorer and more distant _ '' as pointed out by @xcite . thus , he confirms `` _ the bulk _ '' of the sample formed by the stars hd 166 , hd 96064 , hd 97334 , hd 116956 , hd 139777 , hd 139813 and hd 141272 ( see his table 1 ) to be an entity on its own . here we discuss the possible existence of the hercules - lyra mg as an independent association using kinematic ( space motion ) , spectroscopic ( lithium abundance ) and photometric ( isochrone fitting ) criteria . a total of 12 possible members ( stars marked with _ a _ in table [ tab : mgs ] ) have been added to the initial sample of @xcite from our catalogue of _ late - type stars possible members of young stellar kinematic groups _ @xcite . the candidates have been chosen by their kinematics assuming a total dispersion of in @xmath7 and @xmath8 , respectively ; that is , an average position of km s@xmath9 has been determined using the stars given by @xcite , and every star in our catalogue in a radius of has been selected . the value of the dispersion has been chosen equal to that of the old castor mg @xcite , coeval of the hercules - lyra association . no restriction in the @xmath10 component has been imposed in this first selection . in table [ tab : mgs ] we summarize the results obtained by us . from the whole sample of 27 candidates , eight stars have been discarded as members by their space motion : hd 25457 , located inside the b4 subgroup ( see fig . [ fig : uv ] ) ; hd 96064 , hd 112733 , hip 67092 , the binary system made up of the f - type star hd 139777 and hd 139813 , and hd 207129 all them with a value in @xmath10 higher than that of the rest of the candidates ( see fig . [ fig : uvw_mgs ] ) ; and hd 113449 , classified as member of the ab dor mg by @xcite ( see table [ tab : mgs ] and @xmath11 [ sec : abdor ] for a more detailed discussion ) and questioned by @xcite because of its relatively high lithium abundance . we have also studied the lithium abundance measured as the equivalent width of the lithium line @xmath126707.8 , @xmath13(li i ) in each one of the candidates . the values of @xmath13(li i ) have been taken from @xcite and compared with those of the members of well - known stellar clusters ( see fig . [ fig : li_mgs ] ) . the results appear to be consistent with an age of for seven candidates . however , several stars ( hd 1466 , hd 17295 , , and hd 82443 ) show an @xmath13(li i ) comparable to that of the members of the pleiades while other five ( hd 37394 , hd 97334b , hd 111395 , hd 116956 and hd 141272 ) are fully depleted or have a value lower than the expected for a member of the hercules - lyra association . for isochrone fitting , we have adopted pre - main sequence models from @xcite . for @xmath14 k , the models systematically underestimate the age when comparing with clusters of known age such as the pleiades and ic 2391 in a @xmath15 vs. @xmath16 diagram @xcite due to the transformation from flux to colour . bearing this in mind , the corrected transformation adopted by @xcite and @xcite for stars cooler than 4000 k has been used in this work . the values of @xmath16 have been taken from the hypparcos catalogue @xcite . the result of comparing the position of the stars with the isochrones in the colour - magnitude diagram ( cmd ) ( fig . [ fig : vi ] ) is again in agreement with and age of . nevertheless , no conclusions can be inferred from the cmd alone since isochrones of more than 80 myr converge for @xmath17 mag . , and ages larger than 300 myr could be adopted . from the combination of the three criteria , the total sample of candidates is reduced to 10 stars with @xmath13(li i ) and position in the cmd compatibles with an age of , which could form the bulk of the hercules - lyra association , and other 15 definitively non members or with a doubtful classification ( table [ tab : mgs ] ) . the members show a deviation from the centre ( ) lower than that of other coeval mgs such as castor and ursa major ( see * ? ? ? * and references therein ) . a similar dispersion ( ) is found in w , confirming the results in @xmath7 and @xmath8 . in the same way , the shape of the mg in the velocity field is in agreement with the theory of the mgs @xcite . according to this theory , not - gravitational bounded stars formed in the same forming region and with low sigmas in @xmath7 , @xmath8 and @xmath10 are dispersed during their rotation around the galactic centre , inducing a particular shape in both the space and the velocity field since some of the stars fall behind while others go ahead . the galactic potential maintains the group bounded during several hundreds of years , in spite of the initial velocity dispersion in the molecular cloud , in both the @xmath3-plane and the @xmath10 component . very recently , @xcite have identified a large group of stars with the same space motion than the well - known young k - dwarf ab dor ( d = 15 pc ) , a quadruple system @xcite made up of three late - type stars ab dor a ( hd 36705 ) , ab dor ba and ab dor bb and a very low mass companion which has recently been object of discussion because of the discrepancy between its dynamical mass and that predicted by evolutionary models @xcite . all the stars listed in table 1 of @xcite are situated inside the local association ( see fig . [ fig : uv ] ) near the boundaries of the young disk stellar population @xcite , and have at least one indicator of youth . taking the intensity of the h@xmath18 emission line of these stars and the position in a @xmath19 diagram of three m - type members of the mg into account , they estimate an age of @xmath20 myr for the ab dor mg . very recently , @xcite and @xcite have showed that the components of ab dor should have an age of 75 150 myr based on the comparison of both their position in the @xmath21 vs. @xmath19 diagram with respect to the pleiades and ic 2391 clusters , and the @xmath13(li i ) of ab dor a with that of rapidly rotating k dwarfs in the pleiades . moreover , with an age of @xmath4 100 myr the discrepancy between observations and models for the very low - mass companion ( ab dor c ) would disappear ( eg . * ? ? ? taking this into account , they propose an age range of 75 150 myr for all the mg . to the initial sample of @xcite , we have added 13 stars ( marked with _ b _ in table [ tab : mgs ] ) from our catalogue of _ late - type stars possible members of young stellar kinematic groups _ @xcite . these stars have been included to both searches : a ) for other members of the group and b ) to show the existence of two subgroups of different ages more clearly . they have been chosen because of their kinematics , assuming a total dispersion of in @xmath7 and @xmath8 respectively , around the centre of the ab dor mg defined by @xcite . we have imposed no restriction to the @xmath10 component for this first selection . the whole sample contains a total of 50 stars . we have compared the @xmath13(li i ) of every star in the sample with that of known members of young open clusters ( fig . [ fig : li_mgs ] ) , as well as their position in the @xmath16 diagram with the isochrones of @xcite ( fig . [ fig : vi ] ) . the results reveal the existence of two subgroups with stars showing different spectroscopic and photometric features , mixed in the velocity field ( see fig . [ fig : uvw_mgs ] ) . the members of the first subgroup , that includes ab dor and pw and a very active young k2-dwarf @xcite show @xmath13(li i ) similar to that of the high - rotators in the pleiades ( upper continuous line in fig . [ fig : li_mgs ] ) which are above the values found in the low - rotators of ic 2602 ( lower dashed line in fig . [ fig : li_mgs ] ) . their position between the 30 and the 80 myr isochrones in the @xmath16 diagram , together with the first result , is compatible with an age of 30 - 50 myr . moreover , the stars from the sample of @xcite belonging to this subgroup are situated above the sequence of the pleiades in the @xmath21 vs. @xmath19 diagram in @xcite . here we have obtained a dispersion @xmath22 km s@xmath9 in the w component , quite similar to the one observed in other young stellar associations such as tucana or @xmath2 cha ( see * ? ? ? * and references therein ) . for determining the dispersion we have rejected the stars bd+07 1919a and b ( marked with * in table [ tab : mgs ] ) since their radial velocities used for calculating the galactic velocity components have not been corrected for binarity since no orbital solution has been found in the literature . nevertheless , although the membership of this system is not completely reliable taking the value of their @xmath10 component into account , it has been included in the sample as possible member because of the position of the a component in the cmd , which suggests an age of @xmath4 30 myr . the stars in the second subgroup show features , in terms of @xmath13(li i ) values and position in the cmd , comparable with that of the members of the pleiades cluster : in fig . [ fig : li_mgs ] they are situated slightly above the lower envelope of the pleiades , while in fig . [ fig : vi ] they situate on the ( zams ) 80 myrs isochrone . its members could be considered as part of subgroup b4 , one of the four subgroups found by @xcite inside the local association in their study of the space motion of ob associations using hypparcos astrometric data . the authors find a mean age of @xmath4 150 myr for this subgroup using information from the photometry . the higher dispersion found for the stars of this second subgroup in the velocity space ( fig . [ fig : uvw_mgs ] ) is in agreement with the age estimated by us . on the other hand , the results about ab dor mg indicate that this quadruple system has indeed @xmath4 50 myr . the value of @xmath13(li i ) for ab dor a is somewhat above the upper envelope of the pleiades but not so high as the one of ic 2602 ( fig . [ fig : li_mgs ] ) . on the other hand , its ( @xmath16 ) colour situates it between the 30 and 80 myr isochrones ( fig . [ fig : vi ] ) . the same result is clearly visible in fig . 1 of @xcite , where ab dor is situated above the lower sequence of the pleiades in the @xmath21 vs. @xmath19 diagram . with an age of 50 myr , the discrepancy between observations and models for the ab dor very low - mass companion ( ab dor c ) shown in @xcite continuous , although it can be solved if the very low - mass companion were indeed an unresolved binary system @xcite . in table [ tab : mgs ] we list the stars belonging to the nearest moving groups : hercules - lyra association and ab dor mg , and those being part of the local association b4 subgroup . for the hercules - lyra association , a division between certain members and candidates with doubtful classification or non members has been made . in the three groups , new candidates from our catalogue of _ late - type stars possible members of young stellar kinematic groups _ @xcite have been selected because of their kinematics ( see @xmath11 [ sec : herc ] and @xmath11 [ sec : abdor ] ) . a total of 75 stars including the known members and the new candidates selected by us have been analysed . kinematic , spectroscopic and photometric criteria have been utilized to discriminate non members from the rest of candidates of the hercules - lyra association , and to distinguish between the members of the ab dor mg and those of the b4 subgroup . in the velocity space , hercules - lyra is clearly distinguishable from the rest of the sample ( see figs . [ fig : uv ] and [ fig : uvw_mgs ] ) . the dispersion in @xmath7 , @xmath8 , and @xmath10 is comparable with that of other coeval mgs such as castor and ursa major ( e.g. * ? ? ? * ) , and compatible with its age ( see @xmath11 [ sec : herc ] ) . on the other hand , ab dor mg and b4 subgroup are mixed up and age - dating criteria are necessary to distinguish between the members of both groups . nevertheless , the dispersion in @xmath10 for ab dor mg is quite smaller than the one of b4 subgroup . age - dating criteria are also necessary to discriminate non members of hercules - lyra from the certain ones . the results of applying them are summarized in table [ tab : mgs ] : the hercules - lyra association is formed by 10 certain members situated at a mean distance of @xmath4 20 pc and show values of @xmath13(li i ) ( fig . [ fig : li_mgs ] ) and a position in the @xmath16 cmd ( fig . [ fig : vi ] ) compatible with an age of 150 300 myr ; the members of ab dor mg are situated at a mean distance of @xmath4 30 pc and show lithium abundances typical of stars with 30 50 myr ( fig . [ fig : li_mgs ] ) , which is in agreement with their position in the @xmath23 vs. @xmath16 diagram ( fig . [ fig : vi ] ) ; finally , a set of stars with @xmath13(li i ) and positions in the cmd compatible with an age of 80 120 myr are mixed with hercules - lyra and ab dor mg , and have been classified as other members of the local association b4 subgroup ( see @xmath11 3 ) . note that the age estimated using the position of the members of hercules - lyra in the cmd is a lower limit since the 80 myrs isochrone is overlapped with the zams for spectral types earlier than about k5 . on the other hand , the age estimated using the equivalent width of the li i line @xmath126707.8 is more robust since the 50% of the stars classified as members have measurements of the @xmath13(li i ) : the li indicator is useful only for spectral type later than g0 , but only three of the 25 candidates of the initial sample are f stars . stars in these three subgroups form an excellent list of young cool stars for studying how planets are formed , since they cover a range of ages between 30 and 200 myr , characteristic of the period during which the solar system was formed , and they are close enough to be accessible to adaptive optics . in addition , they can be taken as targets for direct imaging detection of sub - stellar companions brown dwarfs and extra - solar giant planets @xcite and for cold dust , debris disks @xcite . this work was supported by the universidad complutense de madrid and the spanish ministerio de educacin y ciencia ( mec ) , programa nacional de astronoma y astrofsica under grants aya2004 - 03749 and aya2005 - 02750 . icc acknowledges support from mec under ap2001 - 0475 . we would like to thanks the referee for useful comments which have contributed to improve the manuscript . agekyan , t. a. , & belozerova , m. a. 1979 , the evaporation of stars from clusters , with the development of coronas and moving clusters , soviet astr . , 23 , 4 ( also astron zh . 56 , 9 ) asiain , r. , figueras , f. , torra , j. , & chen , b. 1999a , , 341 , 427 asiain , r. , figueras , f. , & torra , j. 1999b , , 350 , 434 chen , c. h. , et al . 2005 , , 634 , 1372 close , l. m. , et al . 2005 , nature , 433 , 286 de simone , r. s. , wu , x. , & tremaine , s. 2004 , , 350 , 627 eggen , o. j. 1965 , the observatory , 85 , 191 eggen , o. j. 1975 , , 87 , 37 eggen , o. j. 1983a , , 204 , 377 eggen , o. j. 1983b , , 204 , 391 eggen , o. j. 1984 , , 55 , 597 esa 1997 , the hypparcos and tycho catalogues , esa - sp , 1200 famaey , b. , jorissen , a. , luri , x. , mayor , m. , udry , s. , dejonghe , h. , & turon , c. 2005 , , 430 , 165 fuhrmann , k. 2004 , astronomische nachrichten , 325 , 3 gaidos , e. j. 1998 , , 110 , 1259 gaidos , e. j. , koresko , c. 2004 , new astronomy , 9 , 33 guirado , j. c. , et al . 2006 , , 446 , 733 heartyi , t. , neuhuser , r. , stelzer , b. , fernndez , m. , alcal , j. m. , covino , e. , & hambaryan v. 2000 , , 353 , 1044 jeffries , r. d. 1995 , , 273 , 559 king , j. r. , villarreal , a. r. , soderblom , d. r. , gulliver , a. f. , & adelman , s. j. 2003 , , 125 , 1980 liu , m. c. , et al . 2004 , , 608 , 526 lpez - santiago , j. , montes , d. , fernndez - figueroa , m. j. , & ramsey , l. w. 2003 , , 411 , 489 lpez - santiago , j. 2005 , ph.d . thesis , universidad complutense de madrid ( ucm ) lpez - santiago , j. , montes , d. , glvez , m. c. , crespo - chacn , i. , fernndez - figueroa , m. j. , de castro , e. , & cornide , m. 2005 , , in preparation lowrance , p. , et al . 2005 , , 130 , 1845 luhman , k. l. , & potter 2006 , , in press ( astro - ph0510666 ) luhman , k. l. , stauffer , j. r. , & mamajek , e. e. 2005 , , 628 , l69 marois , c. , macintosh , b. , song , i. , barman , t. 2005 , nature `` brief communications arising '' , astro - ph/0502382 martn , e. l. 2003 , , 126 , 918 masciadri , e. , et al . 2005 , , 625 , 1004 metchev , s. a. , hillenbrand , l. a. , meyer , m. r. 2004 , , 600 , 435 montes , d. , lpez - santiago , j. , glvez , m. c. , fernndez - figueroa , m. j. , de castro , e. , & cornide , m. 2001 , mnras , 328 , 45 neuhuser , r. , brandner , w. , eckart , a. , et al . 2000 , , 354 , l9 ortega , v. g. , de la reza , r. , jilinsky , e. , & bazzanella , b. 2004 , , 609 , 243 quast , g. r. , torres , c. a. o. , de la reza , r. , da silva , l. , & drake , n. 2001 , in asp conf . 244 , young stars near earth : progress and prospects , ed . r. jayawardhana & t. greene ( san francisco : asp ) , 49 quillen , a. c. , minchev , i. 2005 , , 130 , 576 siess , l. , duffour , e. , & forestini , m. 2000 , , 358 , 593 skuljan , j. 1999 , a study of moving groups of stars in our galaxy , ph.d thesis weinberger , a. j. , rich , r. m. , becklin , e. e. , zuckerman , b. , & matthews , k. 2000 , , 544 , 937 wichmann , r. , schmitt , j. h. m. m. , & hubrig , s. 2003 , , 399 , 983 wielen , r. 1971 , , 13 , 309 zuckerman , b. , song , i. , bessell , m. s. , & webb , r. a. 2001 , , 562 , l87 zuckerman , b. , & song , i. 2004 , ara&a , 42 , 685 zuckerman , b. , song , i. , bessell , m. s. 2004 , , 613 , l65 lccccccccccc + 166@xmath24 & 00 06 36.78 & 29 01 17.41 & k0 v & 13.7 & -6.9 0.2 & -15.0 & -21.6 & -10.0 & 0.75 & 0.80 & 75 + 10008 & 01 37 35.47 & -06 45 37.52 & g5 v & 23.6 & 11.6 0.6 & -13.2 & -18.1 & -11.1 & 0.80 & 0.84 & 103 + 233153@xmath25 & 05 41 30.73 & 53 29 23.28 & m0.5 & 12.5 & 1.9 1.0 & -14.4 & -22.9 & -14.3 & 1.40 & 1.91 & 16 + hip 37288@xmath25 & 07 39 23.04 & 02 11 01.18 & k7 & 14.9 & 18.5 5.0 & -11.0 & -21.5 & -13.1 & 1.38 & 1.81 & - + 70573@xmath25 & 08 22 49.95 & 01 51 33.55 & g6 v & 45.7 & 19.5 1.0 & -14.7 & -18.8 & -6.7 & 0.59 & - & 149 + hip 53020@xmath25 & 10 50 52.06 & 06 48 29.34 & m4 & 12.9 & -2.0 0.1 & -7.9 & -22.5 & -19.1 & 1.68 & 2.81 & - + gj 560b@xmath25 & 14 42 30.42 & -64 58 30.50 & k5 v & 16.4 & 7.0 4.0 & -10.9 & -19.2 & -10.8 & 1.15 & - & - + 139664@xmath25 & 15 41 11.38 & -44 39 40.34 & f5 v & 17.5 & -5.4 2.0 & -15.1 & -19.8 & -9.7 & 0.41 & 0.47 & - + 206860@xmath24 & 21 44 31.33 & 14 46 18.98 & g0 v & 18.4 & -16.9 2.0 & -14.6 & -21.4 & -11.0 & 0.58 & 0.66 & 115 + 213845@xmath25 & 22 34 41.64 & -20 42 29.56 & f7 v & 22.7 & -1.9 0.9 & -15.1 & -20.6 & -12.9 & 0.45 & 0.49 & - + + 1466@xmath25 & 00 18 26.12 & -63 28 38.97 & f8 v & 40.9 & 0.54 2.0 & -8.8 & -20.0 & -1.2 & 0.54 & 0.61 & 125 + 17925 & 02 52 32.13 & -12 46 10.97 & k1 v & 10.4 & 17.5 0.1 & -15.0 & -21.8 & -8.7 & 0.88 & 0.91 & 212 + 1e 0318.5 - 19.4@xmath25&03 20 49.50 & -19 16 10.00 & k7 v & 27.0 & 20.8 1.0 & -12.7 & -17.3 & -11.8 & - & - & 63 + 37394 & 05 41 20.34 & 53 28 51.81 & k1 v & 12.2 & 0.3 0.2 & -12.9 & -23.3 & -14.5 & 0.84 & 0.88 & 2 + 82443@xmath24 & 09 32 43.76 & 26 59 18.71 & k0 v & 17.7 & 8.1 0.1 & -9.9 & -22.8 & -5.6 & 0.78 & 0.78 & 176 + 96064 & 11 04 41.47 & -04 13 15.91 & g8 v & 24.4 & 18.3 0.8 & -14.2 & -26.7 & -0.6 & 0.77 & 0.81 & 114 + 97334b@xmath24 & 11 12 32.35 & 35 48 50.69 & g0 v & 21.7 & -3.6 1.0 & -15.8 & -23.2 & -11.2 & 0.60 & 0.67 & 10 + 111395 & 12 48 47.05 & 24 50 24.81 & g5 v & 17.7 & -8.6 1.0 & -18.4 & -21.6 & -9.2 & 0.70 & 0.74 & 0 + 112733@xmath25 & 12 58 31.97 & 38 16 43.55 & k0 v & 22.5 & -3.4 0.1 & -17.6 & -23.3 & -0.8 & 0.74 & 0.79 & 93 + 116956 & 13 25 45.53 & 56 58 13.77 & g9 v & 21.8 & -13.1 0.3 & -15.9 & -18.8 & -8.8 & 0.80 & 0.83 & 0 + hip 67092@xmath25 & 13 45 05.33 & -04 37 13.25 & k5 & 25.7 & 4.6 0.5 & -8.0 & -22.4 & 1.8 & 1.49 & 1.57 & - + 139777 & 15 29 11.20 & 80 26 55.00 & f0 v & 22.1 & -15.8 0.5 & -14.7 & -26.6 & -2.2 & - & - & - + 139813@xmath26 & 15 29 23.60 & 80 27 01.00 & g5 v & 21.7 & -15.8 0.5 & -14.7 & -26.6 & -2.2 & 0.80 & 0.83 & - + 141272 & 15 48 09.46 & 01 34 18.26 & g8 v & 21.4 & -27.2 0.3 & -19.6 & -27.6 & -14.0 & 0.80 & 0.84 & 6 + 207129@xmath25 & 21 48 15.75 & -47 18 13.01 & g0 v & 15.6 & -6.5 1.3 & -13.3 & -22.2 & 0.3 & 0.60 & 0.66 & - + + 1405 & 00 18 20.90 & 30 57 22.03 & k2 v & 30.6 & -11.2 0.1 & -5.3 & -28.8 & -17.8 & 1.04 & - & 268 + hip 6276 & 01 20 32.27 & -11 28 03.74 & ( g8 ) & 35.1 & 9.9 1.0 & -4.7 & -27.8 & -13.6 & 0.79 & 0.83 & 145 + 13482 & 02 12 15.41 & 23 57 29.54 & k1 + k5 & 32.3 & -1.3 0.3 & -7.1 & -28.3 & -11.8 & 1.13 & 1.22 & 110 + 17332 & 02 47 27.42 & 19 22 18.56 & g0 + g5 & 32.6 & 4.1 1.3 & -8.5 & -28.5 & -12.9 & 0.68 & 0.74 & 155 + 19668@xmath27 & 03 09 42.29 & -09 34 46.59 & g0 v & 40.2 & 14.6 0.7 & -5.1 & -28.8 & -10.3 & 0.81 & 0.84 & 191 + 21845 & 03 33 13.49 & 46 15 26.54 & ( g5 ) & 33.8 & -6.0 0.3 & -6.5 & -25.6 & -15.7 & 0.70 & 0.81 & 200 + hip 16563b & 03 33 14.00 & 46 15 19.00 & m0 & 33.8 & -6.1 1.1 & -6.5 & -25.6 & -15.7 & - & - & 30 + 25457 & 04 02 36.74 & -00 16 08.12 & f5 v & 19.2 & 17.0 0.3 & -7.2 & -28.6 & -11.6 & 0.52 & 0.58 & 100 + 25953 & 04 06 41.53 & 01 41 02.08 & f5 & 55.3 & 17.6 0.6 & -6.9 & -27.8 & -14.3 & 0.48 & 0.55 & 120 + 36705 & 05 28 44.83 & -65 26 54.85 & k1 & 14.9 & 33.0 3.0 & -7.5 & -29.8 & -16.0 & 0.83 & 0.94 & 295 + 37572 & 05 36 56.85 & -47 57 52.87 & k0 v & 23.9 & 31.0 1.0 & -7.2 & -26.9 & -13.9 & 0.84 & 0.86 & 240 + bd+20 1790@xmath27 & 07 23 44.00 & 20 25 06.00 & k5 v & 31.6 & 19.9 0.1 & -4.9 & -32.5 & -18.5 & 1.07 & - & 85 + 89744@xmath27 & 10 22 10.56 & 41 13 46.31 & f7 v & 39.0 & -6.5 1.3 & -10.4 & -29.6 & -14.2 & 0.53 & 0.60 & - + 139751 & 15 40 28.39 & -18 41 46.19 & ( k7 ) & 42.6 & -8.9 0.4 & -7.5 & -31.9 & -15.6 & 1.24 & 1.40 & 110 + 160934 & 17 38 39.63 & 61 14 16.12 & m0 & 24.0 & -35.6 0.7 & -5.3 & -27.6 & -14.5 & 1.30 & 2.58 & 40 + hip 106231 & 21 31 01.71 & 23 20 07.37 & k8 & 25.1 & & -5.7 & -27.3 & -15.0 & 1.03 & 1.10 & 233 + 217343 & 23 00 19.29 & -26 09 13.50 & g3 v & 32.1 & 6.3 1.5 & -3.1 & -26.8 & -14.1 & 0.65 & 0.72 & 167 + 217379 & 23 00 27.96 & -26 18 42.80 & ( k8 ) & 30.0 & 8.4 1.5 & -2.0 & -24.5 & -15.4 & 1.34 & 1.59 & 0 + hip 114066 & 23 06 04.84 & 63 55 34.36 & ( m1 ) & 24.9 & -23.7 0.8 & -6.8 & -27.3 & -15.9 & 1.21 & 1.77 & 30 + 218860 & 23 11 52.05 & -45 08 10.63 & ( g5 ) & 50.5 & 10.3 1.2 & -7.9 & -29.1 & -11.3 & 0.71 & 0.76 & 220 + hip 115162 & 23 19 39.56 & 42 15 09.82 & ( g4 ) & 49.4 & -19.7 0.2 & -4.8 & -27.5 & -14.3 & 0.75 & 0.79 & 160 + + 4277 & 00 45 50.89 & 54 58 40.17 & f8 v + k3 & 48.5 & -15.4 0.5 & -8.9 & -26.6 & -15.8 & 0.52 & 0.59 & 119 + 6569 & 01 06 26.15 & -14 17 47.11 & k1 v & 50.0 & 6.0 1.2 & -8.6 & -31.1 & -9.3 & 0.91 & 0.95 & 150 + hip 12635 & 02 42 20.97 & 38 37 21.20 & ( k3.5 ) & 49.6 & -4.1 0.3 & -8.7 & -28.7 & -13.1 & 0.88 & 0.89 & 146 + 16760 & 02 42 21.31 & 38 37 07.20 & g5 & 49.6 & -3.3 0.2 & -9.3 & -28.3 & -13.4 & 0.71 & 0.77 & 158 + hip 14807 & 03 11 12.34 & 22 25 22.77 & ( k6 ) & 49.8 & 4.1 0.3 & -5.1 & -28.6 & -16.1 & - & - & 34 + hip 14809 & 03 11 13.84 & 22 24 57.11 & g5 & 49.8 & 5.2 0.2 & -6.0 & -28.3 & -16.7 & 0.71 & 0.66 & 145 + hip 17695 & 03 47 23.35 & -01 58 19.93 & m3 & 16.3 & 16.0 1.7 & -7.4 & -27.1 & -10.6 & 1.51 & 2.16 & - + 25457 & 04 02 36.74 & 00 16 08.13 & f6 v & 19.2 & 15.3 0.4 & -6.0 & -28.3 & -10.5 & 0.52 & 0.58 & 118 + 35650 & 05 24 30.17 & -38 58 10.76 & ( k7 ) & 17.7 & 30.9 1.0 & -7.1 & -27.0 & -14.5 & 1.25 & 1.21 & 0 + hip 26369 & 05 36 55.07 & -47 57 47.99 & ( k7 ) & 23.9 & 31.1 1.1 & -7.2 & -27.0 & -13.9 & 1.17 & 1.13 & 30 + 45270 & 06 22 30.94 & -60 13 07.15 & g1 v & 23.5 & 30.0 0.7 & -7.6 & -26.7 & -13.6 & 0.61 & 0.66 & 135 + gsc 8894 - 426 & 06 25 55.39 & -60 03 29.20 & m2 & ( 22 ) & 31.8 2.0 & -10.3 & -27.7 & -15.6 & - & - & 0 + 48189 & 06 38 00.36 & -61 32 00.19 & g1.5 v & 21.7 & 33.4 1.0 & -7.1 & -28.5 & -15.0 & 0.62 & 0.69 & 120 + hip 31878 & 06 39 50.02 & -61 28 41.52 & ( k7 ) & 21.9 & 30.5 0.7 & -7.2 & -27.4 & -13.9 & 1.26 & 1.53 & 50 + hip 36349@xmath27 & 07 28 51.37 & -30 14 48.54 & ( m3 ) & 15.6 & 26.6 1.0 & -7.4 & -24.4 & -15.7 & 1.44 & 1.78 & 0 + bd+07 1919a@xmath28 & 08 07 09.09 & 07 23 00.13 & k5 v & 40.2 & 19.1 0.1 & -4.6 & -31.9 & -4.5 & 1.24 & 1.40 & - + bd+07 1919b@xmath28 & 08 07 08.78 & 07 22 58.39 & k7 v & 40.2 & 18.7 0.1 & -7.8 & -24.4 & -1.2 & 1.15 & - & - + hip 51317@xmath27 & 10 28 55.55 & 00 50 27.58 & m2 & 26.1 & 8.3 0.5 & -7.8 & -29.2 & -15.1 & 1.50 & 2.26 & - + 92945@xmath27 & 10 43 28.27 & -29 03 51.43 & k1 v & 21.6 & 23.2 0.6 & -14.9 & -28.3 & -4.0 & 0.87 & 0.92 & 158 + gj 466@xmath27 & 12 25 58.58 & 08 03 44.03 & m0 v & 39.9 & 10.0 0.1 & -9.8 & -32.3 & 0.1 & 1.46 & 1.47 & - + 113449 & 13 03 49.66 & -05 09 42.52 & ( k1 ) & 22.1 & 2.0 0.5 & -5.0 & -28.8 & -9.8 & 0.85 & 0.89 & 142 + 129333@xmath27 & 14 39 00.22 & 64 17 29.84 & g1.5 v & 33.9 & -20.6 0.3 & -7.2 & -29.1 & -4.6 & 0.63 & 0.69 & 189 + hip 81084 & 16 33 41.61 & -09 33 11.95 & m0.5 & 31.9 & -15.0 0.4 & -7.0 & -28.7 & -13.4 & 1.44 & 1.77 & 0 + 152555 & 16 54 08.14 & -04 20 24.66 & g0 & 47.6 & -17.1 0.5 & -6.1 & -28.8 & -12.6 & 0.59 & 0.66 & 133 + 199065a@xmath27 & 20 57 22.44 & -59 04 33.46 & g2 v & 50.9 & 11.0 2.0 & -7.1 & -28.5 & -12.0 & 0.66 & 0.72 & 65 + 199065b@xmath27 & 20 57 21.86 & -59 04 34.23 & g5 v & 50.9 & 11.0 2.0 & -4.6 & -29.7 & -8.6 & - & - & 65 + gj 856a & 22 23 29.09 & 32 27 33.47 & m0 v & 16.1 & -24.0 3.0 & -6.6 & -30.9 & -13.9 & 1.57 & 2.12 & - + gj 856b@xmath27 & 22 23 30.00 & 32 27 00.00 & m1 v & 16.1 & -21.7 1.0 & -6.6 & -28.8 & -14.7 & 1.49 & - & - + 224228 & 23 56 10.67 & -39 03 08.40 & k3 v & 22.1 & 12.1 0.5 & -7.6 & -27.9 & -12.3 & 0.97 & 1.02 & 76 +
the latest results in the research of forming planetary systems have led several authors to compile a sample of candidates for searching for planets in the vicinity of the sun . young stellar associations are indeed excellent laboratories for this study , but some of them are not close enough to allow the detection of planets through adaptive optics techniques . however , the existence of very close young moving groups can solve this problem . here we have compiled the members of the nearest young moving groups , as well as a list of new candidates from our catalogue of _ late - type stars possible members of young stellar kinematic groups _ , studying their membership through spectroscopic and photometric criteria .
during the last decade , spectroscopic and photometric studies have provided strong evidence for multiple stellar populations in a number of galactic globular clusters . in contrast to the standard formation scenario , according to which globular clusters are simple stellar populations composed of stars all with the same age and chemical composition , these studies have revealed star - to - star variations in the abundances of light elements in main sequence and red giant stars , suggesting a second stellar generation ( hereafter sg ) that must have formed out of gas processed through a first generation ( hereafter fg ) of stars . sg stars are identified by their anomalous abundances of na , o , mg and al relative to normal halo field stars of similar metallicity ( see e.g. gratton et al . 2004 , carretta et al . 2009a , b ) . accurate photometric studies of a few globular clusters have strengthened the evidence for multiple populations by showing their fingerprints ( multiple main sequences and sub - giant branches ) in the color - magnitude diagram ( see e.g. lee et al . 1999 , pancino et al . 2000 , bedin et al . 2004 , piotto et al . 2007 , milone et al . 2008 , marino et al . 2008 , anderson et al . the most striking main sequence anomalies are populations with helium abundances much larger than the standard big - bang value expected to be universally present in the oldest stars . the helium overabundance is interpreted as an evolutionary by - product of the progenitors of the anomalous sg stars . the leading models for the source gas from which sg stars formed involve rapidly rotating massive stars ( e.g. decressin et al . 2007 ; see also de mink et al . 2009 for a more recent work suggesting massive binaries as a possible polluting source ) , or intermediate - mass asymptotic giant branch ( agb ) stars ( e.g. ventura et al . 2001 ; see also renzini 2008 for a review and references therein ) . prior to the discovery of the multiple main sequences , the presence of stars with helium anomalies had been predicted based on the very extended horizontal branches of some clusters ( dantona et al . 2002 , dantona and caloi 2004 ) . analysis of horizontal branch morphologies of several clusters also led to the first suggestion that the second stellar generation could account for a high fraction from 30% to 100%of the total number of stars ( dantona & caloi 2008 ) . while the high - quality photometric data necessary to detect multiple populations have been so far obtained for only a few clusters , a recent spectroscopic survey of about 2000 stars in 19 galactic globular clusters has shown evidence of multiple populations in all the clusters observed , and in all cases sg stars account for a significant fraction ( @xmath5 ) of the cluster mass ( carretta et al . 2009a ) . dercole _ et al . _ ( 2008 ) studied the formation and dynamical evolution of multiple populations in globular clusters , and found that two distinct evolutionary processes can cause stars to escape and populate the galactic stellar halo during cluster dynamical evolution . at early times , stars escape over the tidal boundary as the cluster expands due to the loss of primordial gas and fg sn ejecta . since sg stars form preferentially in the central regions of the cluster , almost all of the escapers at this stage are fg stars . the result is that the sg fraction among bound cluster stars increases dramatically . as the cluster evolution continues into the phase dominated by two - body relaxation , the two populations mix and the fractional escape rates of the fg and sg stars due to evaporation tend to equalize , stabilizing the sg fraction ( see also decressin et al . 2008 for another study of the mixing process ) . thus the vast majority of escapers from the cluster are fg stars . two recent spectroscopic studies ( carretta et al . 2010 , martell & grebel 2010 ) have found that the vast majority of halo stars studied have abundances typical of fg stars in clusters ; only about 1.52.5% of the stars are na - rich ( carretta et al . 2010 ) and cn - strong ( martell & grebel 2010 ) , and hence classifiable as sg stars . many general studies have addressed the evolution of the galactic globular cluster system and the dispersal of cluster stars in the halo ( see e.g. kroupa & boily 2002 , baumgardt et al . 2008 , portegies zwart et al . 2010 , vesperini 2010 and references therein ) . however , this letter is specifically focused on sg stars . our goals here are to estimate both the number of sg stars that may have formed in the galactic globular cluster system and the fraction , @xmath1 , of sg stars formed in globular clusters that now populate the galactic stellar halo , and to illustrate the link between sg formation , globular cluster evolution and the formation history of the galactic halo . in [ sec : hydro ] , we present the results of hydrodynamical simulations exploring the dependence of the mass of sg stars formed on cluster initial properties . in [ sec : sghalo ] we introduce our numerical framework to estimate the fraction of sg stars in the halo , and discuss its theoretical and observational ingredients . in [ sec : res ] we present and discuss our results . in order to calculate the total mass of sg stars formed in a cluster as a function of cluster initial properties , we have carried out a series of one - dimensional hydrodynamical simulations modeling the sg formation process . the technical details of our simulations are described in dercole _ et al . _ we outline here only the aspects relevant to this letter . we assume that the fg cluster initially follows a king ( 1962 ) density profile @xmath6^{-1.5}\ ] ] out to some truncation radius @xmath7 . stellar masses are distributed in the range @xmath8 according to either a kroupa et al . ( 1993 ; hereafter k93 ) or a kroupa ( 2001 ; hereafter k01 ) stellar imf . we model the formation of sg stars from the ejecta of agb stars and assume that the sg formation phase extends from about 30 myr after the fg cluster forms until @xmath9 myr ( should the sg star formation epoch extend to @xmath10 myr , and hence include ejecta of stars down to 4 @xmath11 , the sg mass can be obtained by multiplying the values presented below by a factor @xmath12 ) . as discussed in dercole _ et al . _ ( 2008 ) and dercole _ et al . _ ( 2010 ) in order to explain the chemical properties of sg stars it is necessary to invoke the presence of pristine gas mixed in with the matter from which sg stars formed ( see also pflamm - altenburg & kroupa 2009 for a study of the possible accretion of gas from the ism during the sg formation process ) . as shown by dercole et al . ( 2008 , 2010 ) , a mass of pristine gas equal to about 50% of the total mass of sg star - forming gas is required to reproduce the abundance patterns ( na - o , mg - al anticorrelations and the helium distribution function ) observed in the massive cluster ngc2808 and in the low - mass cluster m4 ( see also ventura _ et al . _ 2009 , ventura & dantona 2009 ) . however , additional systematic studies of the chemical properties of the individual clusters for which spectroscopic data are available will be required to explore the allowed mass range of pristine gas and any possible dependences on protocluster conditions . here we simply assume that pristine material makes up 50% of the total mass of gas forming the sg stars . results for different pristine gas fractions can be easily obtained from those presented in this letter . the two panels of fig . [ fig : hydrofig ] show the fractional mass of sg stars formed , @xmath13 , as a function of the total cluster mass ( @xmath14 ) , for models with different combinations of values of the half - mass radius ( @xmath15 pc ) , of the truncation radius ( @xmath16 , 40 , and 80 pc ; these values encompass the range of r for the majority of galactic globular clusters , for which the modal value is @xmath17 pc ; see e.g. mackey & van den bergh 2005 ) and for a k93 or k01 imf . as the cluster mass increases for given values of @xmath7 and @xmath18 , so too do the cluster escape speed , the fraction of agb ejecta retained , and the total mass of sg stars formed . for the structural parameters explored , only clusters initially more massive than @xmath19 can retain enough agb ejecta to form an appreciable fraction of sg stars . the difference between the values of @xmath20 obtained for the k93 and the k01 imf is a consequence of the fact that the k01 imf contains a larger fraction of the cluster mass in the mass range of the polluting agb stars . a complete model for the determination of the fraction of sg stars in the galactic halo would require accurate knowledge of the initial global properties of the galactic globular cluster system as well as its dynamical history , the initial internal structure and subsequent dynamical evolution of clusters with multiple stellar populations , and a comprehensive theory of how the stellar halo formed . much work remains to be done on all of these problems . considering the uncertainties involved in all of these key ingredients , we elect to proceed here in a simplified way , and our estimation of @xmath2proceeds as follows . for a given value of the total mass of the galactic globular cluster system , @xmath21 , and a given initial globular cluster system mass function ( hereafter igcmf ) , we calculate the total mass of sg stars formed in the cluster system , @xmath22 , using the results of the hydrodynamical simulations described in the previous section . @xmath22 may be written , in general , as an integral over cluster structural properties @xmath23 where @xmath24 is the joint distribution function of initial cluster mass and structural parameters . we assume either a schechter igcmf ( see e.g. burkert & smith 2000 ) @xmath25 or a simple power - law ( e.g. whitmore 2003 ) @xmath26 with these assumptions , and for given values of @xmath18 and @xmath7 at the time of sg formation ( @xmath27 myr ) , eq.[eq : msg1 ] becomes @xmath28 neglecting mass loss due to stellar evolution , the total mass of the stellar halo in the form of sg stars , either from dissolved globular clusters or evaporated from those remaining , is the difference between @xmath22 and the current total mass of sg stars still in clusters , @xmath29 . a systematic observational study of sg stars in galactic globular clusters is still lacking , and will require significant effort . currently the only study affording an estimate of the fraction of sg stars in a significant number of galactic clusters , as well as exploration of possible trends of the sg fraction with other cluster properties , is the spectroscopic survey of @xmath30 stars in 19 galactic globular clusters carried out by carretta _ ( 2009a , b ) . they find that all clusters studied host sg stars , and that the fraction of individual cluster masses in sg stars ranges from about 50% to more than 80% . hence we define the current cluster sg mass fraction @xmath31 by @xmath32 where @xmath33 is the current total mass of halo galactic globular clusters ( assuming @xmath34 ; mackey & van den bergh 2005 ) . most of the clusters studied are more massive than @xmath35 , and it is not known whether such a large sg fraction is common to all galactic globular clusters or only to those currently more massive than some threshold . for purposes of this letter we simply assume as a reference value @xmath36 . results for different values of @xmath31 can be very easily derived from those shown in this letter . to complete our calculation we need an estimate of @xmath21 , the initial total mass of all stars in globular clusters . this is not easily determined from observations of the current galactic globular cluster system . here again , we simply parametrize the normalization of eq.[eq : msg0 ] or [ eq : msg2 ] by defining the parameter @xmath37 where we take @xmath38 ( see e.g. freeman & bland - hawthorn 2002 ) . the current fraction of sg stars in the stellar halo , @xmath1 , then is @xmath39 finally , if we define @xmath40 as the ratio of the total initial mass of sg stars in the galactic globular cluster system to the total initial mass of the cluster system @xmath41 and use the definition of @xmath31 in eq.[eq : msgnow ] along with the values of @xmath42 and @xmath43 adopted above , we can rewrite eq.[eq : fsg ] as @xmath44 the form of eq.[eq : fsgb ] allows to easily identify the different theoretical and observational ingredients needed for the determination of @xmath1 . we summarize them here . * @xmath40 , the ratio of the total initial mass of sg stars to the total initial mass of the galactic globular cluster system , incorporates the dependency on the sg formation model ( through the function @xmath45 ) , as well as on the stellar imf , the cluster mass and structural parameter distribution functions ( see eq.[eq : msg ] above ) . * @xmath46 , the ratio of the total initial mass of the galactic globular cluster system to the current halo mass , is not known . the simple linear scaling of @xmath2with this parameter allows to easily explore the dependence of @xmath2on the total initial mass of the galactic globular cluster system . * @xmath31 , the fraction of the total current mass of the halo cluster population composed of sg stars , is a parameter to be determined observationally . in our study , based on the spectroscopic survey of carretta _ et al._(2009a , b ) , we adopt as a reference value @xmath47 . however , in this case too , the simple linear scaling of @xmath2with @xmath48allows to easily calculate @xmath2for other values of @xmath48 . in fig.[fig : fsg1 ] we show the dependence of @xmath40 on the power - law index @xmath49 of the igcmf . different lines correspond to different pairs of values of @xmath18 and @xmath7 , a k93 or a k01 imf and different values for the parameters of the igcmf ( @xmath50 and @xmath51 ; @xmath52 is fixed at @xmath53 ) . an increase in @xmath50 and/or @xmath51 increases the fraction of the total initial mass in clusters massive enough to form sg stars ( @xmath54 ; see fig . [ fig : hydrofig ] ) , and hence increases @xmath40 . similarly , as @xmath49 decreases the fraction of total mass in massive clusters increases , and so does @xmath40 . for a fixed value of @xmath7 , decreasing the value of @xmath18 adopted has the obvious consequence of increasing the fraction of agb ejecta retained and the amount of sg stars formed in lower - mass clusters . the absolute upper limit ( @xmath55 for a k93 imf and @xmath56 for a k01 imf ) would be attained by assuming ( unrealistically ) that all clusters ( including the low - mass ones ) were compact enough at the time of sg formation ( @xmath57 myr ) to retain all the agb ejecta and reach a fraction of sg stars formed @xmath58 or @xmath59 for a k93 and a k01 imf respectively . from the results shown in fig.[fig : fsg1 ] one can easily calculate @xmath2as a function of @xmath46 and @xmath48for any of the combinations of igcmf , cluster structural parameters and imf explored . for example , fig.[fig : fsg2 ] , shows the dependence of @xmath2on @xmath46 for a power - law igcmf with @xmath60 , for various values of @xmath51 , @xmath7 , @xmath18 and for either a k93 or a k01 imf ( and , as discussed above , we have adopted @xmath47 ) . if clusters were at the time of sg formation ( @xmath57 myr ) more compact than assumed here , the lines of @xmath2vs @xmath46 would fall between those shown in fig.[fig : fsg2 ] and that corresponding to the absolute upper limit easily obtained using eq.[eq : fsgb ] and the maximum value of @xmath40 ( @xmath61 for a k93 imf and @xmath62 for a k01 imf ) . figs . [ fig : fsg1 ] and [ fig : fsg2 ] demonstrate that , for the broad range of values of the igcmf parameters considered , the predicted fraction of sg stars in the halo is always small , @xmath3 for a k93 imf and @xmath4 for a k01 imf . this result is consistent with the small fraction ( @xmath01.5 - 2.5% ) of halo stars identified as sg stars by carretta et al . ( 2010 , based on the stars sodium abundance ) and martell and grebel ( 2010 , based on cn and ch bandstrengths ) . for comparison , 50% of globular cluster stars are identified as sg when the same classification criteria are adopted ( kraft 1994 , carretta et al . 2010 , martell & grebel 2010 ) . we note that for these estimates of @xmath1 , using , for example , a power - law igcmf with @xmath60 as a reference model , our calculations ( fig.[fig : fsg2 ] ) with a k01 imf ( k93 imf ) imply that a fraction ranging from about 20% to about 40% ( 30% to about 60% ) of the galactic stellar halo must be composed of stars originally formed in globular clusters . although this estimate is very uncertain and both more calculations and further observational studies of sg stars in the halo will be required to refine it , it is interesting to note how the fractions of sg stars in globular clusters and in the galactic stellar halo connect and constrain both models for the formation and dynamical evolution of multiple stellar generations in globular clusters and the formation history of the galactic halo . * acknowledgments . * ev and sm were supported in part by nasa grants nnx07ag95 g , nnx08ah15 g and nnx10ad86 g and by nsf grant ast-0708299 . ad acknowledges financial support from italian miur through grant prin 2007 ( prot . 2007jjc53x ) . fd was supported by prin miur 2007 multiple stellar populations in globular clusters : census , characterization and origin ( prot . n. 20075tp5k9 ) .
many observational studies have revealed the presence of multiple stellar generations in galactic globular clusters . these studies suggest that second - generation stars make up a significant fraction of the current mass of globular clusters , with the second - generation mass fraction ranging from @xmath050 to 80% in individual clusters . in this letter we carry out hydrodynamical simulations to explore the dependence of the mass of second - generation stars on the initial mass and structural parameters and stellar initial mass function ( imf ) of the parent cluster . we then use the results of these simulations to estimate the fraction @xmath1 of the mass of the galactic stellar halo composed of second - generation stars that originated in globular clusters . we study the dependence of @xmath2on the parameters of the initial mass function of the galactic globular cluster system . for a broad range of initial conditions , we find that the fraction of mass of the galactic stellar halo in second - generation stars is always small , @xmath3 for a kroupa-1993 imf and @xmath4 for a kroupa-2001 imf .
the field of fracture mechanics is becoming extremely broad with the occurrence of unexpected failure of weapons , buildings , bridges , ships , trains , airplanes , and various machines . there are two fundamental fracture criterions : the strain energy release rate ( i.e. , g theory ) and the stress intensity factor ( i.e. , k theory ) ; see arthur and richard ( 2002 ) . the experimentally determined stress intensity depends on the specimen size , and the fracture is accompanied by energy localization and concentration . + localization is manifested by degradation of material properties with localized large deformations , and this feature often results in formation and propagation of macrocracks through engineering structures . due to the importance of localization phenomena in structural safety assessment , much research has been conducted to resolve experimental , theoretical and computational issues associated with localization problems , as reviewed by chen and schreyer ( 1994 ) and chen and sulsky ( 1995 ) . for hyperelastic materials , important progress has been made based on the gradient approach ; see aifantis ( 1984 ) and triantafyllidis and aifantis ( 1986 ) . however , there is still a lack of analytical results for three - dimensional boundary - value problems . in this paper , hence , we study the localization in a slender cylinder composed of an incompressible hyperelastic material subjected to tension , based on an analytical approach to solve the three - dimensional governing equations . we also intend to provide mathematical descriptions for some interesting phenomena as observed in experiments . + jansen and shah ( 1997 ) conducted careful experiments on concrete cylinders by using the feedback - control method . from two test series , the typical stress - displacement behavior for different height - diameter ratios with normal strength and high strength was obtained . it appears that the pre - peak segment of the stress - displacement curves agrees well with the pre - peak part of the stress - strain curves , but the post - peak segment shows a strong dependence on the geometric size ( i.e. , the radius - length ratio ) . more specifically , the longer the specimen is , the steeper the post - peak part of the stress - displacement curves becomes . also , they found that the width of the localization zone changes with the specimen size . in the experiment by gopalaratnam and shah ( 1985 ) , it was found that the tangent value in the ascending part of the stress - strain curves seemed to be independent of the specimen size but in the post - peak part there was a softening region and no unique stress - strain relation . schreyer and chen ( 1986 ) studied the softening phenomena analytically based on a one - dimensional model . their results indicate that if the size of the softening zone is small enough ( in a relative sense ) , the behavior of displacement - prescribed loading is unstable , and the softening curves are steeper than those with a larger size of the softening region . here , we shall provide the three - dimensional analytical solutions to capture all the localized features mentioned above . + another purpose of this research is to provide a method judging the onset of failure in a slender cylinder subjected to tension . here , we use the maximum - distortion - energy theory ( the huber - hencky - von mises theory ; see riley et al . 2007 ) , which depicts there are two portions of the strain energy intensity . one is the portion producing volume change which is ineffective in causing failure by yielding , and the other is that producing the change of shape which is completely responsible for the material failure by yielding . + by constructing the analytical solutions for localizations , it is possible to get the point - wise energy distribution . then , an expression for the maximum value of the strain energy can be obtained . with the huber - hencky - von mises theory , we can then establish an analytical criterion for identifying the onset of failure . + mathematically , to deduce the analytical solutions for localizations in a three - dimensional setting is a very difficult task . one needs to deal with coupled nonlinear partial differential equations together with complicated boundary conditions . further , the existence of multiple solutions ( corresponding to no unique stress - strain relation ) makes the problem even harder to solve . here , the analysis is carried out by a novel method developed earlier ( dai and huo , 2002 ; dai and fan , 2004 ; dai and cai , 2006 ) , which is capable of treating the bifurcations of nonlinear partial differential equations . our results yield the analytical forms of the strain and stress fields , total elongation , the potential energy distribution and the strain energy distribution , which are characterized by localization phenomena . in particular , it is found that once the localization is formed its width does not change with the further increase of the total elongation , which is in agreement with the experimental observations . we also provide a description for the snap - through phenomenon . + the remaining parts of the paper are arranged as follows . in section 2 , we formulate the three - dimensional govering equations for the axisymmetric deformations of a circular cylinder . we nondimensionalize them in section 3 to identify the key small variable and key small parameters . then , in section 4 , a novel method of coupled series - asymptotic expansions is used to derive the normal form equation of the original system . by the variational principle , in section 5 , we derive the same equation by considering the energy . in section 6 , with the help of the phase plane , we solve the boundary - value problems for a given external axial force and a given elongation , respectively . in section 7 , through an energy analysis , we determine the most preferred configurations and give a description of the snap - through phenomenon . also , an analytical criterion for identifying material failure based on the huber - hencky - von mises theory is discussed . finally , concluding remarks and future tasks are given in section 8 . we consider the axisymmetric deformations of a slender cylinder subjected to a static axial force at one plane end with the other plane end fixed . the radius of the cylinder is @xmath0 and the length is @xmath1 . we take a cylindrical coordinate system , and denote @xmath2 and @xmath3 as the coordinates of a material point of the cylinder in the reference and current configurations , respectively . the radial and axial displacements can be written as @xmath4 we suppose that the cylinder is composed of an incompressible hyperelastic material , for which the strain energy density @xmath5 is a function of the first two invariants @xmath6 and @xmath7 of the left cauchy - green strain tensor , i.e. , @xmath8 . moreover the first piola - kirchhoff stress tensor @xmath9 is given by @xmath10 where @xmath11 is the deformation gradient and @xmath12 is the indeterminate pressure . if the strains are small , it is possible to expand the first piola - kirchhoff stress components in terms of the strains up to any order . the expressions for the stress components are very lengthy , and due to the complexity of calculations , we shall only work up to the third - order material nonlinearity . the formula containing terms up to the third - order material nonlinearity has been provided by fu and ogden ( 1999 ) : @xmath13 where + @xmath14 + @xmath15 is the pressure value in eq . ( 2 ) corresponding to zero strains , @xmath16 is the incremental pressure , and @xmath17 and @xmath18 are incremental elastic moduli defined by @xmath19 it can be found that @xmath20 where @xmath21 denotes the first - order partial derivative of @xmath5 with respect to the invariant @xmath7 at @xmath22 , @xmath23 denotes the first - order partial derivative of @xmath5 with respect to the invariant @xmath6 at @xmath22 . in the following derivations , we shall also use @xmath24 to denote the i - th order and the j - th order partial derivative of @xmath5 with respect to the invariants @xmath6 and @xmath7 at @xmath22 . all the coefficients in eq . ( [ axishu ] ) can be expressed in terms of @xmath25 and @xmath26 and here for brevity the expressions are omitted . + the equilibrium equations for a static and axisymmetric problem are given by @xmath27 @xmath28 the incompressibility condition yields that @xmath29 we consider the case where the lateral surface of the cylinder is traction - free , then the stress components @xmath30 and @xmath31 should vanish on the lateral surface . thus , we have the boundary conditions : @xmath32 eqs . ( [ e1])([t1 ] ) together with eq . ( [ ji ] ) provide three governing equations for three unknowns @xmath33 and @xmath16 . the former two are very complicated nonlinear partial differential equations ( pde s ) and the boundary conditions ( [ b1 ] ) are also complicated nonlinear relations ( cf . ( [ 12])([16 ] ) ) . to describe the localization , one needs to study the bifurcation of this complicated system of nonlinear pde s . as far as we know , there is no available mathematical method . here , we shall adapt a novel approach involving coupled series - asymptotic expansions to tackle this bifurcation problem . a similar methodology has been developed to study nonlinear waves and phase transitions ( dai and huo , 2002 ; dai and fan , 2004 ; dai and cai , 2006 ) . first , we shall nondimensionalize this system to identify the relevant small variable and small parameters . we introduce the dimensionless quantities through the following scales : @xmath34 where @xmath1 is the length of the cylinder , @xmath35 is a characteristic axial displacement and @xmath36 is the material shear modulus , with a transformation being defined by @xmath37 substituting eqs . ( [ 2 ] ) , ( [ 10 ] ) and ( [ 11 ] ) into eqs . ( [ e1])([t1 ] ) , we obtain @xmath38 @xmath39 @xmath40+\varepsilon^2[v^2w_z+2sv(v_sw_z - v_zw_s)]=0 , \end{array}\label{14}\ ] ] where @xmath41 is regarded as a small parameter . for convenience , we have replaced @xmath42 by @xmath43 in the non - dimensionalized equations . here and thereafter , a subscript letter is used to represent the corresponding partial derivative ( i.e. , @xmath44 ) . the full forms of ( [ 12 ] ) and ( [ 13 ] ) are very lengthy and we do not write out the nonlinear terms explicitly for brevity . + substituting ( [ 10 ] ) and ( [ 11 ] ) into the traction - free boundary conditions ( [ b1 ] ) , we have @xmath45+\varepsilon^2[-p^*v^2 + 3b_{47}v^3 + 30b_7v^2w_z+b_{16}vw_z^2 + 4b_8w_z^3\\\vspace{.2 cm } + \nu(-2p^*(2vv_s+v_zw_s)+b_{22}v^2v_s+7b_9vv_z^2 + b_{23}vv_zw_s\\\vspace{.2 cm } + 28b_9vw_s^2 + 72b_{12}vv_sw_z+2b_{19}v_z^2w_z+2b_8v_zw_sw_z+8b_{19}w_s^2w_z\\\vspace{.2 cm } + b_{18}v_sw_z^2)+\nu^2(-4p^*v_s^2 + 3b_{31}vv_s^2 + 4b_{19}v_sv_z^2 + 8b_{13}v_sv_zw_s\\+16b_{19}v_sw_s^2+b_{20}v_s^2w_z)+\nu^3b_{21}v_s^3]|_{s=\nu}=0 , \end{array}\label{15}\ ] ] @xmath46+\varepsilon^2[-p^*(vv_z+v_zw_z)\\\vspace{.2 cm } + b_{44}v^2v_z+2b_{16}v^2w_s+b_{30}vv_zw_z+b_{32}vw_sw_z+b_{13}v_zw_z^2\\\vspace{.2 cm } + b_{33}w_sw_z^2+\nu(-2p^*v_sv_z+b_{23}vv_sv_z - b_{3}v_z^3 + 56b_9vv_sw_s\\\vspace{.2 cm } + 2b_{11}v_z^2w_s-12b_3v_zw_s^2 + 16b_5w_s^3 + 2b_8v_zv_sw_z\\ + 16b_{19}v_sw_sw_z)+\nu^2(4b_{13}v_s^2v_z+16b_{14}v_s^2w_s)]|_{s=\nu}=0 , \end{array}\label{16}\ ] ] where @xmath47 is a small parameter for a slender cylinder . + then , eqs . ( [ 12])([16 ] ) compose a new system of complicated nonlinear pde s with complicated boundary conditions , which is still very difficult to solve exactly . however , it is characterized by a small variable @xmath48 and two small parameters ( @xmath49 and @xmath50 ) , which permit us to use expansion methods to proceed further . + * remark : * the coefficients @xmath51 in eqs . ( [ 12])([16 ] ) can be expressed in terms of @xmath52 and @xmath53 , and for brevity we omit their expressions . we note that @xmath48 is also a small variable as @xmath54 . an important feature of the system ( [ 12])([16 ] ) is that the unknowns @xmath55 and @xmath16 become the functions of the variable @xmath56 , the small variable @xmath48 and the small parameters @xmath49 and @xmath50 , i.e. , @xmath57 to go further , we assume that @xmath58 have the following taylor expansions in the neighborhood of the small variable @xmath59 : @xmath60 substituting eqs . ( [ 18])([20 ] ) into eq . ( [ 13 ] ) and equating the coefficient of @xmath61 to be zero yields that @xmath62 similarly , substituting eqs . ( [ 18])([20 ] ) into eq . ( [ 12 ] ) and setting the coefficients of @xmath61 and @xmath63 to be zero , we obtain @xmath64 @xmath65 the expressions of @xmath66 and @xmath67 are very lengthy , whose expressions are omitted for brevity . from the incompressibility condition ( [ 14 ] ) , the vanishing of the coefficients of @xmath61 and @xmath63 leads to the following two equations : @xmath68 @xmath69 substituting eqs . ( [ 18])([20 ] ) into the traction - free boundary conditions ( [ 15 ] ) and ( [ 16 ] ) , and neglecting the terms higher than @xmath70 , we obtain @xmath71 @xmath72 where the lengthy expressions for @xmath73 are omitted for brevity . ( [ 21])([27 ] ) are seven nonlinear ordinary differential equations , which are the governing equations for the seven unknowns @xmath74 and @xmath75 . mathematically , it is still very difficult to solve them directly . to go further , we shall use the smallness of the parameter @xmath49 . from eq . ( [ 24 ] ) , we obtain @xmath76 using the above equation in eq . ( [ 25 ] ) , we can express @xmath77 in terms of @xmath78 and @xmath79 , and then from eqs . ( [ 21 ] ) and ( [ 23 ] ) , we can express @xmath80 and @xmath75 in terms of @xmath81 and @xmath82 . substituting the expressions for @xmath83 and @xmath75 into eq . ( [ 22 ] ) , @xmath84 can be expressed in terms of @xmath78 and @xmath79 . then , from eq . ( [ 26 ] ) , @xmath82 can be expressed in terms of @xmath78 and @xmath79 . finally , from eqs . ( [ 22 ] ) and ( [ 27 ] ) , we obtain @xmath85 @xmath86 we note that the above two equations come from the axial equilibrium equation ( the coefficient of @xmath61 ) and the zero tangential force at the lateral surface , the two most important relations for tension problems in a slender cylinder . + by eliminating @xmath87 from eqs . ( [ 29 ] ) and ( [ 30 ] ) and then expressing @xmath79 in terms of @xmath78 , finally we obtain an equation for the single unknown @xmath78 : @xmath88 by further using eq . ( [ 28 ] ) , we obtain the following equation for the axial strain @xmath89 : @xmath90 integrating eq . ( [ 36 ] ) once , we obtain @xmath91 where @xmath92 is an integration constant . to find the physical meaning of @xmath92 , we consider the resultant force @xmath93 acting on the material cross section that is planar and perpendicular to the cylinder axis in the reference configuration , and the formula is + @xmath94 after expressing @xmath95 in terms of @xmath89 by using the results obtained above , the integration can be carried out , and as a result we find that @xmath96 comparing eqs . ( [ jifen ] ) and ( [ tzhi ] ) , we have @xmath97 . thus , we can rewrite eq . ( [ tzhi ] ) as @xmath98 if we retain the original dimensional variable and let @xmath99 we have @xmath100 where @xmath101 since eq . ( [ 41 ] ) is derived from the three - dimensional field equations , once its solution is found , the three - dimensional strain and stress fields can also be found . also , it contains all the required terms to yield the leading - order behavior of the original system . therefore , we refer eq . ( [ 41 ] ) as the normal form equation of the system of nonlinear pde s ( [ 12])([14 ] ) together with boundary conditions ( [ 15 ] ) and ( [ 16 ] ) under a given axial resultant . it is also possible to deduce the equation for @xmath102 by considering the total potential energy and then using the variational principle . by using the expansions obtained in section 4 , we can express the two principal invariants @xmath6 and @xmath7 in terms of @xmath89 . the results are @xmath103,\\\vspace{0.3 cm } i_2 - 3=\displaystyle3\varepsilon^2w_{0z}^2 - 4\varepsilon^3w_{0z}^3 + 5\varepsilon^4w_{0z}^4+s[\varepsilon^2(\frac{81}{64}w_{0zz}^2-\frac{15}{8}w_{0z}w_{0zzz})\\ \displaystyle+\varepsilon^3(-\frac{3(53\phi_{01}+71\phi_{10})}{64(\phi_{01}+\phi_{10})}w_{0z}w_{0zz}^2 + \frac{3(25\phi_{01}+16\phi_{10})}{8(\phi_{01}+\phi_{10})}w_{0z}^2w_{0zzz})].\label{43 } \end{array}\ ] ] from eqs . ( [ 43 ] ) , we know the first terms in the right - hand sides of eqs . ( [ 43 ] ) are second - order nonlinear . to be consistent with the third - order material nonlinearity of the stress components , the strain energy @xmath5 should be expanded up to the fourth - order nonlinear terms . so , according to the taylor s expansion , we have @xmath104+\cdots.\label{44 } \end{array}\ ] ] in eq . ( [ 44 ] ) it is sufficient to keep the second - order terms of @xmath105 and @xmath106 . substituting the expressions of @xmath6 and @xmath7 in eqs . ( [ 43 ] ) into eq . ( [ 44 ] ) , we have @xmath107.\label{45 } \end{array}\ ] ] the stored energy per unit length is given by @xmath108 substituting eq . ( [ 45 ] ) into eq . ( [ 46 ] ) and carrying out the integration , we obtain the average stored energy over a cross section : @xmath109.\label{47 } \end{array}\ ] ] letting @xmath99 we can further rewrite the above equation as @xmath110,\label{48 } \end{array}\ ] ] where @xmath111 is the young s modulus , @xmath112 and @xmath113 are constants related to material parameters . + the total potential energy for a force - controlled problem is given by @xmath114 by the variational principle , we have the following euler - lagrange equation : @xmath115 which yields that @xmath116 which is just eq . ( [ 41 ] ) . this shows that the normal form equation ( [ 41 ] ) obeys the variational principle for energy . + if we multiply @xmath117 to both sides of eq . ( [ 52 ] ) , it can be integrated once to yield that @xmath118 where @xmath119 is an integration constant . + in the following section , we shall discuss the solutions for two boundary - value problems based on eqs . ( [ 52 ] ) and ( [ 53 ] ) , and reveal their main characteristics . we rewrite eq.([52 ] ) as a first - order system as follows : @xmath120 without loss of generality , we take the length @xmath1 of the cylinder to be @xmath121 , then @xmath0 is equivalent to the radius - length ratio . we suppose that the two plane ends of the cylinder are attached to rigid bodies . then we have @xmath122 and@xmath123 we point out that although eq . ( [ 52 ] ) is one - dimensional , it is derived from the three - dimensional governing equations , and as a result we can also derive the proper boundary conditions by considering the condition in the other ( radial ) dimension such as eq . ( [ 58 ] ) . if one directly introduces a one - dimensional model ( say , using a gradient theory ) , such an option is not available . so , this is another advantage of eq . ( [ 52 ] ) . + from eqs . ( [ 57 ] ) and ( [ 58 ] ) , we have @xmath124 substituting eq . ( [ 59 ] ) into eq . ( [ t1 ] ) and integrating with respect to @xmath125 once , we obtain @xmath126 thus , the proper boundary conditions for eq . ( [ 52 ] ) are @xmath127 to solve this boundary - value problem of the first - order system ( [ 56 ] ) under eq . ( [ 61 ] ) , we shall conduct a phase - plane analysis with the engineering stress as the bifurcation parameter . the critical points of this system is given by @xmath128 here we shall consider a class of strain energy functions @xmath129 such that the @xmath130 plot based on eq . @xmath131 has one peak , and this requires that @xmath132 the @xmath130 curves corresponding to eq . @xmath131 are plotted in fig . 1 . in this figure , @xmath133 is the local maximum of the stress and @xmath134 is the corresponding strain value , and they are given by @xmath135 when we take @xmath136 and @xmath137 , @xmath138 ; when we take @xmath139 and @xmath137 , @xmath140 ; when we take @xmath141 and @xmath137 , @xmath142 . the three curves in fig . 1 correspond to these values of @xmath143 , @xmath144 and @xmath145 , respectively . + in the following discussions we consider the tension case so that @xmath146 . similar analysis can be made for the compression case , which will not be discussed here . equation ( [ 53 ] ) can be rewritten as @xmath147 in this paper , we consider the case of @xmath148 . new phenomena can arise for @xmath149 and the results will be reported elsewhere . for the present case , the phase plane always has a saddle point and a center point as @xmath150 varies , which is shown in fig . 2 . in this figure , @xmath151 and @xmath152 are a saddle point and a center point , respectively . there are two solutions for the same stress @xmath150 , which are represented by the curve 1 and the curve 2 in fig . 2 , respectively . for curve 1 , the right hand of eq . ( [ y59 ] ) have four real roots , which we label in an increasing order by @xmath153 and @xmath154 . we note that the smallest root @xmath155 is smaller than @xmath156 . so , from eq . ( [ y59 ] ) we obtain the following expression : @xmath157 then , an integration leads to @xmath158\\ \displaystyle\frac{1}{2}+\frac{a}{\sqrt{2d_2}}\int_v^{\alpha_1}\sqrt{\frac{1 - 8d_3t}{(t-\alpha_1)(t - g_1)(t - g_2)(t-\alpha_2)}}dt , z\in [ \frac{1}{2},1 ] . \end{array}\right.\label{y61}\ ] ] by eq . ( [ 61 ] ) , @xmath155 can be determined by the following two equations : @xmath159 @xmath160 once @xmath155 is found , the solution corresponding to curve 1 can be obtained from eq . ( [ y61 ] ) by numerical integration . in fig . 3 , we have plotted the solution curves for three different values of the engineering stress @xmath150 . from this figure , we find there is nearly a uniform extension in the middle part , but there are two boundary - layer regions near the two ends of the cylinder in order to satisfy the boundary conditions . + there is another solution which is represented by curve 2 in fig . 2 , and we denote the point as @xmath161 at which @xmath162 . then , eq . ( [ y59 ] ) can be rewritten as @xmath163},\label{64}\ ] ] where @xmath164 is another real root of the right - hand of eq . ( [ y59 ] ) , and @xmath165 and @xmath166 . + then , we obtain @xmath167}}dt , z\in [ 0,\frac{1}{2 } ] \\ \displaystyle\frac{1}{2}+\frac{a}{\sqrt{2d_2}}\int_v^{\alpha}\sqrt{\frac{1 - 8d_3t}{(\alpha - t)(\beta - t)[(t - m)^2+n^2]}}dt , z\in [ \frac{1}{2},1 ] . \end{array}\right.\label{65}\ ] ] by eq . ( [ 61 ] ) , @xmath161 can be determined by @xmath168 @xmath169 by numerical integration , we can get @xmath161 from eqs . ( [ 66 ] ) and ( [ 67 ] ) . then the solution corresponding to curve 2 can be obtained from ( [ 65 ] ) by numerical integration . in fig . 4 , we have plotted the solution curves for three different values of the engineering stress @xmath150 . in this figure , there is a sharp - change region in the middle of the slender cylinder , that represents the localization and energy concentration . moreover , the tip is sharper when the engineering stress is smaller . + from eq . ( [ 65 ] ) , one can see that the localization solution depends on @xmath170 through the form @xmath171 and this implies that the localization zone width is proportional to @xmath0 for a fixed length ; see jansen and shah ( 1997 ) . + the solutions obtained above are for a given @xmath150 . to obtain the solutions for a displacement - controlled problem , we follow the idea in dai and bi ( 2006 ) . for that purpose , we need to get the engineering stress - strain curve . + the total elongation is given by @xmath172 where the total elongation @xmath173 is actually the engineering strain since we have taken the length of the cylinder to be 1 . according to the symmetric phase plane and eqs . ( [ y61 ] ) and ( [ 65 ] ) , @xmath174 is a function of @xmath170 , so we can get the total elongation by numerical integrations . in fig . 5 , we have plotted the curves between the total elongation @xmath173 and the engineering stress @xmath150 with different material coefficients corresponding to fig . 1 . segment 1 corresponds to the solution given by eq . ( [ y61 ] ) ( we call it as solution 1 ) , and segment 2 corresponds to the solution given by eq . ( [ 65 ] ) ( we call it as solution 2 ) . for a displacement - controlled problem ( i.e. , given @xmath173 ) , from fig . 5 , we can get the corresponding @xmath150 value(s ) , then the solution(s ) is given by eq . ( [ y61 ] ) or eq . ( [ 65 ] ) . + from eqs . ( 1 ) , ( [ 10 ] ) , ( [ 11 ] ) , ( [ 19 ] ) and ( [ 20 ] ) , we can get the shapes of the cylinder corresponding to eq . ( [ 65 ] ) under different material coefficients for a given @xmath173 , which are shown in fig . 6 , where we take @xmath175 and @xmath176 . in this figure , the width @xmath177 of the localization is defined as @xmath178 , where @xmath179 is the point where the rate of the slope of the surface radial displacement is the maximum . from the above figure , one can see that for different material coefficients the localization widths are different and the localization width is almost the same for the same material coefficients with different loads of engineering stress . that is to say , for different materials the localizations have different widths , but for the same material , the localization width is almost uniform during the loading process . + here and thereafter , we fix the material constants to be @xmath180 and @xmath137 . by the same way , we can get the relations between the total elongation @xmath173 and the engineering stress @xmath150 with different radius - length ratios , which are shown in fig . 7 . in this figure , we observe that there is a snap - back for a relatively small value of @xmath0 . we also see that the point @xmath181 ( across which there are multiple values for @xmath150 for a given @xmath173 ) moves up and toward right as the value of @xmath0 increases . for example , when @xmath182 and when @xmath183 . the post - peak curves show very significant changes . there is no unique stress - displacement relationship in the post - peak region . the thinner the specimen is , the steeper the curve becomes , which is in agreement with the experimental results by jansen and shah ( 1997 ) . from this figure , we see that for @xmath184 , unstable behavior is predicted for a displacement - controlled loading whereas larger values of @xmath0 yield results that are stable . similar conclusions follow from the examples given by schreyer and chen ( 1986 ) . + as to @xmath185 , there is a stable relation between the total elongation @xmath173 and the engineering stress @xmath150 , which is in agreement with the experiment result by gopalaratnam and shah ( 1985 ) , who conducted tensile tests on concrete under carefully controlled loading conditions and with refined measuring techniques . + we note that for a displacement - controlled problem , after the elongation @xmath186 ( cf . 7 ) there are bifurcations from one solution to two solutions ( at @xmath187 ) , to three solutions ( @xmath188 ) , to two solutions ( @xmath189 ) , and to one solution ( @xmath190 ) . the shapes of the cylinder corresponding to these solutions are shown in fig . 8 for @xmath191 and @xmath184 . the above figure also manifests that the middle of the cylinder becomes thinner than the two ends as we pull the slender cylinder . the middle part is thinner as the engineering stress decreases for a given @xmath173 , which agrees well with the experimental results . as discussed in the previous section , for a relatively small @xmath0 , there are multiple solutions for @xmath186 . of course , in reality only one solution can be observed at one instant . in this section , we shall further consider the energy values for these solutions to deduce which one is most preferred . + from eq . ( [ 53 ] ) , we have @xmath192 substituting eq . ( [ fang66 ] ) into eq . ( [ 52 ] ) , we obtain @xmath193 then by using eq . ( [ 49 ] ) , we can express the potential energy ( for a given @xmath150 ) in terms of @xmath174 ( scaled by @xmath194 ) : @xmath195 the stored energy is given by @xmath196 then from eqs . ( 58 ) , ( 61 ) , ( 67 ) and ( 68 ) , one can calculate the energy distributions for a given elongation . the stored energy curves corresponding to those values of @xmath173 in fig . 8 are plotted in fig . 9 . in this figure , labels 1 , 2 and 3 correspond to different values of @xmath150 ( in a decreasing order ) . + in fig.9(a ) , the total stored energy values for curve 1 and curve 2 are respectively @xmath197 and @xmath198 . thus for a displacement - controlled problem , as @xmath199 , the shape in the right of fig . 8(a ) represents a preferred configuration , and at @xmath187 there could be a bifurcation from solution 1 to solution 2 ( a localization solution ) . correspondingly , there is a snap - through in the engineering stress - strain curve at @xmath187 . + in fig.9(b ) , the total stored energy values for curve 1 , curve 2 and curve 3 are respectively @xmath200 and @xmath201 . for a displacement - controlled problem , as @xmath202 is the smallest , the shape in the bottom of fig . 8(b ) represents a preferred configuration . + in fig.9(c ) , the total stored energy value for curve 1 and curve 2 are respectively @xmath203 and @xmath204 . for a displacement - controlled problem , as @xmath199 , the shape in the right of fig . 8(c ) represents a preferred configuration . + in fig.10 , we have plotted the engineering stress - strain curve corresponding to the preferred configuration for a displacement - controlled problem . we see that a snap - through takes place at @xmath187 , which leads to the localization ( as represented by solution 2 ) . once the localization happens , there is a high concentration of energy around the middle of the cylinder . it is known that if the strain energy density reaches a critical value there will be the material failure . the analytical results obtained here can be used to calculate the stored energy at any material point . the largest energy value is attained at @xmath205 at which @xmath206 ( cf . ( 61)(63 ) ) . from eqs . ( 40 ) , ( 65 ) , and ( 66 ) , we can express the energy value at this point in terms of @xmath161 , and the result is @xmath207 the values of @xmath208 corresponding to those values of @xmath173 ( in an increasing order ) in fig . 8 for preferred configurations are respectively @xmath209 and @xmath210 . + based on the maximum - distortion - energy theory ( the huber - hencky - von mises theory ; see riley et al . 2007 ) , there are two portions of the strain energy intensity : one for volume change and the other for shape change . in the present work , we consider an incompressible material , so there is no strain energy intensity corresponding to the volume change . then the strain energy is only due to distortion . on the other hand , the strain energy intensity attains its maximum value at the material point @xmath211 . thus , we can give the failure criterion @xmath212 where @xmath213 is the failure value of the strain energy intensity for a given material . fracture will occur whenever the energy by eq . ( 69 ) exceeds the limiting value @xmath213 . to study the localization phenomenon , a phenomenological approach is employed to formulate a three - dimensional boundary value problem with an incompressible hyperelastic constitutive law . a coupled series - asymptotic expansion procedure is developed to solve the non - dimensionalized system of governing differential equations with given boundary data for a slender cylinder subjected to axial tension . with the assumptions appropriate for the slender cylinder , analytical solutions have been obtained for the axisymmetric boundary value problem , which demonstrate the essential features of localization problems and are consistent with the experimental data available . specifically , the width of localization zone depends on the material parameters , and it remains unchanged for the same material in the post - peak regime . also , the snap - back and snap - through phenomena could be predicted with the analytical approach , and a preferred configuration in the post - peak regime could be identified via an energy analysis . due to the lack of three - dimensional analytical results available in the open literature , the analytical work presented in this paper could complement the analytical , experimental and numerical efforts made by the research community for the localization problems over the last several decades . + as indicated by buehler et al . ( 2003 ) , the hyperelasticity is crucial for understanding and predicting the dynamics of brittle fracture . especially , the effect of hyperelasticity is important for understanding the failure evolution in nanoscale materials . since localization identifies the onset of material failure , future work will focus on the identification of the parameters proposed in the current phenomenological model , and on the linkage between the continuum and fracture mechanics approaches , via multiscale analysis . the work described in this paper is supported by two grants from cityu ( project nos : 7002107 and 7001861 ) . dai , h. h. , cai , z. x. , 2006 . phase transitions in a slender cylinder composed of an incompressible elastic material . i. asymptotic model equation . ser . a math . 462 , 7595 . dai , h. h. , fan , x. j. , 2004 . asymptotically approximate model equations for weakly nonlinear long waves in compressible elastic rods and their comparisons with other simplified model equations . solids 9 , 6179 .
in this paper , we study the localization phenomena in a slender cylinder composed of an incompressible hyperelastic material subjected to axial tension . we aim to construct the analytical solutions based on a three - dimensional setting and use the analytical results to describe the key features observed in the experiments by others . using a novel approach of coupled series - asymptotic expansions , we derive the normal form equation of the original governing nonlinear partial differential equations . by writing the normal form equation into a first - order dynamical system and with the help of the phase plane , we manage to solve two boundary - value problems analytically . the explicit solution expressions ( in terms of integrals ) are obtained . by analyzing the solutions , we find that the width of the localization zone depends on the material parameters but remains almost unchanged for the same material in the post - peak region . also , it is found that when the radius - length ratio is relatively small there is a snap - back phenomenon . these results are well in agreement with the experimental observations . through an energy analysis , we also deduce the preferred configuration and give a prediction when a snap - through can happen . finally , based on the maximum - energy - distortion theory , an analytical criterion for the onset of material failure is provided . , , localization ; hyperelasticity ; cylinder ; bifurcations of pde s
various modern ( semi)phenomenological nucleon - nucleon ( nn ) potentials @xcite adjusted to the nucleon - nucleon data allow for a relatively good description of few - nucleon systems . moreover , the existing models of a three nucleon force ( 3nf ) when combined with those potentials led generally to an improvement in the description of three and four nucleon bound states @xcite and reactions in the 3n - continuum @xcite . despite indisputable successes such forces lack reliable theoretical justification in the underlying theory of strong interactions - the quantum chromodynamics ( qcd ) . recently , a new family of nuclear potentials has been derived @xcite which has a direct connection to qcd . incorporating the spontaneously and explicitly broken chiral symmetry , which plays a fundamental role in qcd , it is possible to construct nuclear forces in the framework of chiral perturbation theory . this approach has been originally proposed by weinberg @xcite and further developed by ordez et al . @xcite , kaiser et al . @xcite , entem et al . @xcite , and by epelbaum et al . @xcite . ( see also recent review articles @xcite and references therein ) . up to now , nn and 3n forces have been worked out and applied in few - nucleon systems up to next - to - next - to - next - to - leading order ( n@xmath3lo ) and next - to - next - to - leading order ( n@xmath4lo ) , respectively . the leading 4n - forces have also been worked out recently @xcite . the chiral forces have already been successfully used to describe the elastic nucleon - deuteron scattering @xcite and the deuteron breakup processes @xcite . we remind the reader that applications of chiral nuclear forces are limited to the low - energy region . chiral effective field theory has also been used to study processes with external pions and/or photons , see @xcite for some recent applications . while the neutron - proton radiative capture has already been considered in this framework @xcite , no such analysis is available for electromagnetic processes with three nucleons . in this study we would like , at least partially , to fill this gap by presenting for the first time the chiral predictions for the @xmath5 and @xmath6 processes . we use the chiral nn potentials @xcite at next - to leading order ( nlo ) , next - to - next - to leading order ( n@xmath4lo ) and at next - to - next - to - next - to leading order ( n@xmath3lo ) of chiral expansion . in the n@xmath4lo where the 3nf contributes for the first time , we also include such 3nf s @xcite consistent with the two - body interaction at this order . a complete study of electromagnetic processes in the chiral effective field theory would require the usage of the nuclear electromagnetic current operator derived consistently with the applied nuclear forces . such a current operator is , however , not yet available , see ref . @xcite for some early work along these lines . for that reason , we decided to adopt in the present study our standard approach @xcite , in which the electromagnetic current operator is taken in the form of a one - body current supplemented with some many - body contributions introduced via the siegert theorem . in section 2 we shortly describe our theoretical formalism . results and a comparison to data are presented in section 3 . we summarize in section 4 . the present study is based on the formalism described in detail in @xcite . the calculations are done for the two - body photodisintegration of the 3n bound state and the nuclear matrix element for the radiative @xmath7-capture @xmath8 is found by applying the time reversal to the photodisintegration amplitude @xmath9 . the latter one is a matrix element of the electromagnetic current operator @xmath10 between the initial 3n bound state @xmath11 and the final scattering state @xmath12 @xmath13 in the faddeev scheme @xmath9 can be written in the form @xcite @xmath14 where @xmath15 is a faddeev component of the state @xmath12 and @xmath16 is a permutation operator defined as a sum of cyclical and anti - cyclical permutations of three particles @xmath17 where @xmath18 interchanges the i - th and j - th nucleons . starting from the schrdinger equation for the scattering state @xmath12 the transition amplitude can be written as @xmath19 where @xmath20 is a product of the internal deuteron state and a momentum eigenstate of the free nucleon - deuteron motion . the auxiliary state @xmath21 fulfills the faddeev - like equation @xmath22 where @xmath23 is a part of the 3nf which is symmetrical under the exchange of nucleons 2 and 3 , @xmath24 is the free three - nucleon propagator , and @xmath25 is the two - body t - operator acting in the 2 - 3 subspace . equation ( [ eq : utilde ] ) is solved in the momentum - space basis @xmath26 where @xmath27 and @xmath28 are the magnitudes of the two jacobi momenta . the quantum numbers @xmath29 , @xmath30 , and @xmath31 are the two - body subsystem orbital angular momentum , spin and the total angular momentum , respectively . the orbital angular momentum @xmath32 of the spectator nucleon together with its spin @xmath33 couples to the total spectator angular momentum @xmath34 . the angular momenta @xmath31 and @xmath34 couple finally to the total angular momentum of the 3n system @xmath35 . similar coupling @xmath36 occurs for the isospin quantum numbers . starting from the transition matrix element @xmath37 the observables follow in a standard way @xcite . the chiral potentials depend on low energy constants ( lec ) and the cut - off parameters @xmath38 and @xmath39 . the lec s are determined for given @xmath38 and @xmath39 from a fit to nn and 3n data . the @xmath38 cut - off appears in the spectral function regularization @xcite and ensures that the high - momentum components of the two - pion exchange are explicitly excluded from the potential . the cut - off @xmath39 comes from a further regularization of the potential as used in the lippmann - schwinger equation , which cuts off the ( meaningless ) contributions of high - momentum states in the dynamical equation . we performed numerical calculations for a few combinations of @xmath40 values which covers the possible ranges of the regularization parameters : ( 450(400 at nlo ) , 500 ) , ( 600(550 at nlo),500(600 at n@xmath3lo ) ) , ( 550,600 ) , ( 450(400 at nlo),700 ) , ( 600(550 at nlo),700 ) ( in units of @xmath41 $ ] ) . the theoretical predictions at a given chiral order provide a band comprising all results . the bands presented in the following correspond to the nn potential at nlo , n@xmath4lo , nn+3nf at n@xmath4lo , and finally to the nn potential at n@xmath3lo . let us start the presentation of our results with the c.m . differential cross section for the neutron - deuteron ( figs . [ fig1a]-[fig2 ] ) and proton - deuteron ( figs . [ fig3a]-[fig4 ] ) capture . in fig . [ fig1a ] we display predictions at incoming neutron lab . energy @xmath42=9 mev and 10.8 mev . the light ( cyan ) band corresponds to nlo predictions while the dark ( red ) one refers to the full ( i.e. nn+3nf ) n@xmath4lo results . while the nlo band is relatively broad , the n@xmath4lo band is much narrower . this shows the fast convergence of the chiral expansion for this observable . the nn+3nf n@xmath4lo chiral predictions are in agreement with the av18+urbana ix combination of the standard nn and 3nf models obtained using the same current operator . in fig . [ fig1 ] we show a scale of 3nf effects at n@xmath4lo by comparison predictions which corresponds to the full hamiltonian ( dark band ) and other ones which corresponds to the interaction restricted to the nn force only ( light band ) . results obtained with nn force at n@xmath4lo are similar to those at nlo . due to the width of the nn band it is difficult to conclude if the inclusion of the 3nf improves or not the description of the data , however the 3nf effects are clear . there is an improvement at smaller @xmath43 angles but at larger angles the data favour the pure nn predictions . the comparison between n@xmath4lo and n@xmath3lo based on nn interactions only is shown in fig . the inclusion of the additional terms appearing in n@xmath3lo ( without corresponding 3nfs ) shifts the predictions above those of n@xmath4lo and brings them above the data . it would be interesting to see if the inclusion of the 3nfs in n@xmath3lo will decrease the cross section bringing theory back to the data , similarly as it is the case at n@xmath4lo . in figs . [ fig3a]-[fig4 ] proton - deuteron capture cross sections are shown at two incoming deuteron lab . energies , e@xmath44=29 and 95 mev . the nice convergence of chiral expansion is seen at both energies ( see fig . [ fig3a ] ) . in fig . [ fig3 ] , at smaller deuteron energy e@xmath44=29 mev the inclusion of the 3nf in n@xmath4lo shifts the results in the wrong direction , deteriorating the description of data . passing to ( incomplete ) n@xmath3lo improves the agreement with data . at the much higher deuteron energy e@xmath44=95 mev the 3nf acts in the opposite direction , undoubtedly improving the data description . at this energy there is no difference between predictions at n@xmath4lo and n@xmath3lo based on the nn interactions only . however , the band of predictions at n@xmath3lo is much broader than the corresponding one at n@xmath4lo . now let us discuss selected spin observables . we would like to show results for the deuteron vector analyzing power @xmath1 ( figs . [ fig5a]-[fig6 ] ) and the deuteron tensor analyzing power @xmath2 ( figs . [ fig7a]-[fig8 ] ) at four deuteron energies 17.5 mev , 29 mev , 45 mev and 95 mev , for which comparison with data is possible . here , the convergence of the chiral expansion is much slower ( see figs . [ fig5a ] and [ fig7a ] ) and the bands have more or less the same width . at the highest deuteron energy e@xmath44=95 mev for @xmath1 the n@xmath4lo band is even wider than that one in nlo . the effects of the 3nf at n@xmath4lo on @xmath1 are not substantial and are seen mainly at the highest deuteron energy . the observed discrepancy with the data might at least partially be caused by the simplified model of the electromagnetic operator used in the present calculations . ( the results based on the av18+urbana ix potential and a more realistic model of the current operator is shown in @xcite where a better agreement with data was found . ) the effects of the new terms occurring in the nn potential in n@xmath3lo are small , as seen in fig . [ fig6 ] . at the three lower energies they slightly shift predictions in the data direction . at the highest energy the picture for @xmath1 is similar when including the 3nf or the new n@xmath3lo terms . however , in the first case the band of predictions is much broader . similarly to the @xmath1 case , the effects of moving from nlo to n@xmath4lo , as well as the 3nf effects for @xmath2 ( figs . [ fig7a]-[fig7 ] ) are not significant . the discrepancy between av18+urbana ix and the chiral predictions for the higher energies of 45 mev and 95 mev is particularly interesting . at these energies the step to the next order in the chiral expansion seems to be necessary ( see fig . [ fig8 ] ) . moreover , the additional terms in the n@xmath3lo nn chiral potential shift predictions farther away from the data . this points to the need of developing the corresponding 3n forces , which is under way . the first predictions for radiative nucleon - deuteron capture based on chiral forces are presented . the differential cross section and the deuteron vector and tensor analyzing powers have been calculated using the chiral nn potential at nlo , n@xmath4lo and n@xmath3lo . also the results at n@xmath4lo with the chiral 3nf of that order are shown . the nuclear current operator contains , besides the single nucleon current , some many - body contributions as introduced via the siegert theorem . the chiral potentials give similar predictions as the av18+urbana ix combination of the standard nuclear forces . we found , however , a significant theoretical uncertainty for all considered observables . the bands of predictions are in some cases too broad , especially in the n@xmath3lo nn and n@xmath4lo nn+3nf cases , to make a decisive conclusion about the quality of the nd - capture data description . this broadness can be partially caused by the inconsistent current operator used in this study . in future studies the nuclear current constructed in the same framework as chiral forces should be used . this first study indicates that the inclusion of the consistent current operator and going to n3lo in the chiral expansion might be necessary to obtain a more precise description of radiative processes within the chiral effective field theory framework . however , we have shown that already now this framework can be used for analyzing low energy electromagnetic processes . this work has been supported by the polish committee for scientific research under grant no . 2p03b00825 and by the helmholtz association , contract number vh - ng-222 . the numerical calculations have been performed on the the ibm regatta p690 + of the nic in jlich , germany . stoks , r.a.m . klomp , c.p.f . terheggen , j.j . de swart , _ phys . rev . _ * c49 * , 2950 ( 1994 ) . wiringa , v.g.j . stoks , r. schiavilla , _ phys . rev . _ * c51 * , 38 ( 1995 ) . r. machleidt , f. sammarruca , y. song , _ phys . * c53 * , r1483 ( 1996 ) . j. carlson , r. schiavilla , _ rev . of mod . phys . _ * 70 * , 743 ( 1998 ) . pieper , v.r . pandharipande , r.b . wiringa , j. carlson , _ phys . rev . _ * c64 * , 014001 ( 2001 ) . a. nogga , et al . * 85 * , 944 ( 2000 ) . a. nogga , et al . rev . _ * c65 * , 054003 ( 2002 ) . w. glckle , h. witaa , d. hber , h. kamada , j. golak , _ phys . rept . _ * 274 * , 107 ( 1996 ) . e. epelbaum , _ prog . ( in press ) , nucl - th/0509032 . p.f . bedaque and u. van kolck , _ annu . nucl . part . _ * 52 * , 339 ( 2002 ) . d.r . entem and r. machleidt , _ phys . _ * c68 * , 041001 ( 2003 ) . e. epelbaum , w. glckle , u .- g . meiner , _ nucl . phys . _ * a747 * , 362 ( 2005 ) . s. weinberg , _ phys . lett . _ * b251 * , 288 ( 1990 ) ; _ nucl . _ * b363 * , 3 ( 1991 ) ; _ phys . lett . _ * b295 * , 114 ( 1992 ) . c. ordez , u. van kolck , _ phys . * b291 * , 459 ( 1992 ) ; c. ordez , l. ray , u . van kolck , _ phys . rev . _ * c53 * , 2086 ( 1996 ) . n. kaiser , r. brockmann and w. weise , _ nucl . _ * a625 * , 758 ( 1997 ) ; n. kaiser , _ phys . rev . _ * c61 * , 014003 ( 2000 ) ; _ phys . _ * c62 * , 024001 ( 2000 ) ; _ phys . _ * c64 * , 057001 ( 2001 ) ; _ phys . rev . _ * c65 * , 017001 ( 2002 ) . e. epelbaum , w. glckle , u .- g . meiner , _ eur . phys . j. _ * a19 * , 401 ( 2004 ) . beane , p.f . bedaque , m.j . savage and u. van kolck , _ nucl . phys . _ * a700 * , 377 ( 2002 ) . e. epelbaum , _ to be published in phys . b _ , nucl - th/0511025 . e. epelbaum , a. nogga , w. glckle , h. kamada , u .- g . meiner , h. witaa , _ phys . _ * c66 * , 064001 ( 2002 ) . st . kistryn et al . , _ phys . * c72 * , 044006 ( 2005 ) . beane , v. bernard , e. epelbaum , u.g . meiner and d.r . phillips , _ nucl . phys . _ * a720 * , 399 ( 2003 ) . beane , m. malheiro , j.a . mcgovern , d.r . phillips and u. van kolck , _ nucl . _ * a747 * , 311 ( 2005 ) . h. krebs , v. bernard , u .- g . meiner , _ eur . phys . j. _ * a22 * , 503 ( 2004 ) . v. lensky , v. baru , j. haidenbauer , c. hanhart , a.e . kudryavtsev and u.g . meiner , _ eur . phys . j. _ * a26 * , 107 ( 2005 ) . park et al . , _ phys . * b472 * , 232 ( 2000 ) . t.s.park , d.p.min and m.rho , _ nucl . phys . _ * a596 * , 515 ( 1996 ) . _ , _ phys . rev . _ * c62 * , 054005 ( 2000 ) . r. skibiski , j. golak , h. kamada , h. witaa , w. glckle , a. nogga , _ phys . rev . _ * c67 * , 054001 ( 2003 ) . j. golak , r. skibiski , h. witaa , w. glckle , a. nogga , h. kamada , _ phys . rept . _ * 415 * , 89 ( 2005 ) . g. mitev et al . , _ phys . _ * c34 * , 389 ( 1986 ) . belt , c.r . bingham , m.l . halbert , a. van der woude , _ phys . * 24 * , 1120 ( 1970 ) . pitts et al . , _ phys . _ * c37 * , 1 ( 1988 ) .
present day chiral nucleon - nucleon potentials up to next - to - next - to - next - to leading order and three nucleon forces at next - to - next - to leading order are used to analyze nucleon - deuteron radiative capture at deuteron laboratory energies below e@xmath0 100 mev . the differential cross section and the deuteron analyzing powers @xmath1 and @xmath2 are presented and compared to data . the theoretical predictions are obtained in the momentum - space faddeev approach using the nuclear electromagnetic current operator with exchange currents introduced via the siegert theorem . the chiral forces provide the same quality of data description as a combination of the two - nucleon av18 and the three - nucleon urbana ix interactions . however , the different parametrizations of the chiral potentials lead to broad bands of predictions .
high quality spatially resolved measurements with current x - ray satellites have clearly shown that intracluster ( ic ) gas is not isothermal . with increasing image detail it will likely be found that non - isothermality is a general feature of clusters . this is only to be expected , given the gas extent , origin , and the processes that have affected its evolution . the only justification for spectral fitting of the emission from a large cluster region by a single temperature plasma emission code ( what has been a standard procedure ) was insufficient spectral and spatial resolution . not only the assumption of isothermality is generally invalid when characterizing the emission from a large cluster region , the expectation that the emission is _ purely _ thermal may also be doubtful when a wide spectral band is considered . this is particularly so in clusters with known extended regions of radio emission , where the radio emitting relativistic electrons give rise to non - thermal ( nt ) x - ray emission by compton scattering off the cosmic microwave background ( cmb ) radiation . the well known possibility that cluster x - ray spectra may have high energy power law tails ( , rephaeli 1977 ) has been largely ignored , also because of the low sensitivity of past x - ray satellites to the detection of the predicted low - level nt emission . non - isothermal gas distribution and nt emission are of considerable interest both intrinsically , for the detailed understanding of the astrophysics of clusters , as well as in the use of cluster properties and phenomena as cosmological probes . x - ray emission is currently our best probe of the properties of ic gas ; the thermal structure of the gas can yield important information on energy exchange and transport processes . more globally , the detailed gas density and temperature profiles are required in analysis of measurements of the sunyaev - zeldovich ( s - z ) effect and its use as a cosmological probe of the global parameters of the universe and its large scale structure . on the other hand , nt phenomena in clusters have the potential to contribute significantly to our understanding of the origin of relativistic particles and magnetic fields . it is obvious , therefore , that there is strong motivation for a more realistic characterization of cluster x - ray spectra . the search for nt x - ray emission in clusters has been advanced considerably by the rxte and bepposax satellites . we have initiated and analyzed long rxte observations of the coma cluster ( rephaeli , gruber & blanco 1999 , rephaeli & gruber 2002 ) , a2319 ( gruber & rephaeli 2002 ) , and a2256 ( rephaeli & gruber 2003 ) . in all three clusters we found evidence for a second spectral component . while a secondary emission component would generally reflect the non - isothermality of ic gas , we found that in these three clusters the additional emission is likely to be mostly nt in origin . in each of the clusters this conclusion is based on the relative quality of the statistical fits ( of thermal vs. nt models ) , and the high physical plausibility of nt x - ray emission in the central cluster region in which extended radio emission was measured . similar bepposax searches for nt in these and other clusters ( including a119 , a754 , and a2199 ) also yielded evidence for nt emission in at least some of the clusters ( fusco - femiano 1999 , kaastra 1999 ) , fusco - femiano 2000 , fusco - femiano 2003 ) . note that the significance of the bepposax / pds results has been questioned by rossetti & molendi ( 2003 ) , who claim that the level of instrumental error is higher than what was previously assumed , and that the detection of nt emission in coma is much less significant than originally reported by fusco - femiano ( 1999 ) . however , this claim has been disputed by fusco - femiano ( 2004 ) . the moderately distant cluster a2163 , @xmath4 , is one of the hottest , most luminous clusters , and a prime target of radio , s - z , and x - ray observations in recent years . its extended central region of radio emission ( herbig & birkinshaw 1994 ) was extensively mapped by feretti ( 2001 ) ; this radio ` halo ' is one of the largest and most luminous . spatial correlation between bright regions in the _ chandra _ and radio maps may be interpreted as evidence for intense merging activity ( markevitch & vikhlinin 2001 ) , which is thought to enhance the efficiency of particle acceleration . the gas temperature profile and a detailed temperature map were determined from _ xmm _ ( pratt 2001 ) and _ chandra _ ( govoni 2004 ) measurements . here we report the results from a very long observation of a2163 , the most distant of the radio ( ` halo ' ) clusters observed with the rxte . a2163 observations with the proportional counter array ( pca ) and the high energy x - ray timing experiment ( hexte ) on rxte were made during 95 separate pointings in the period august 24 , 2004 - march 11 , 2005 . data from pca detectors 0 and 2 were collected in the ` good xenon ' spectral mode , which produces a 256-channel count spectrum nominally from 2 to 1000kev . the background estimation tool was found to produce spectra with a several percent gain error for the epoch of these observations ; this was corrected to better than 1% in the software . data from the two independent hexte clusters of detectors were taken in event - by - event mode , and were subsequently accumulated into 256-channel spectra spanning 12250kev . to subtract the background , each hexte cluster was commanded to beamswitch every @xmath5 between on - source and alternate off - source positions 1.5@xmath6 on either side . standard screening criteria were applied to the data segments ( earth elevation angle , spacecraft pointing , avoidance of the south atlantic anomaly , times of geomagnetic activity ) , resulting in a net exposure time for pca of @xmath7 ks and an average exposure time of @xmath8 ks for the hexte clusters . the hexte net observation times were shorter than the pca time because hexte spends half of the time measuring the background , and some more time is lost from electronic dead time caused by cosmic rays . the pca background was estimated with the ` l7/240 ' faint source model provided by the instrument team , and was corrected for errors in detector gain independently in the two detectors . uncertainties in the background correction precluded using pca data above 17 kev . the total pca counting rate was 8 count / s ( c / s ) , or 21% of background . for hexte the net counting rate was 0.6 c / s , or 0.4% of background . the pca background level is estimated to be accurate within @xmath00.5% , or 0.2 c / s . the hexte background , which is directly measured , has been determined ( macdonald 2000 ) to be accurate to within a few hundredths c / s in long exposures . inspection of the screened light curves for pca and hexte revealed no significant variation , as expected for a cluster of galaxies . accordingly , we co - added all the selected pca and hexte data to form net spectra for analysis . spectra from the individual pca detectors were also coadded , and the small differences in gain were accounted for in the generation of the energy response matrices , following procedures prescribed by the pca analysis team . a small energy dependence for systematic errors ( e.g. , gruber et al 2001 ) averaging about 0.8% , was applied to the pca spectrum . additionally , pca spectral channels below 3kev and above 17kev were excluded because of sensitivity to artifacts in the background model , as well as the rapidly declining effective area of the pca outside these bounds . upon inspection of the high energy channels of hexte we restricted the analysis to data in the range 1280kev . we fit the joint pca and hexte spectra summed over all observations to three simple spectral models : a raymond - smith ( r - s ) thermal plasma emission model , two r - s models at different temperatures , and a r - s plus a power law model . in addition we have checked whether we can find a meaningful indication for a temperature structure by fitting the data with two r - s thermal components plus a power law . in all cases most of the observed flux is in a primary @xmath9 kev r - s component . best - fit parameters and 90% confidence intervals are listed in table 1 . the best - fit temperature in the single isothermal gas model is @xmath10 kev , in good agreement with the range determined from _ xmm _ ( pratt 2001 ) and _ chandra _ ( govoni 2004 ) . the observed @xmath11 kev fe xxv k@xmath12 line yields an abundance of @xmath13 @xmath14 ( where @xmath14 denotes solar abundance ) , quite consistent with previously determined values ( @xmath15 ) . no cold absorption was measurable , and given the 3 kev pca threshold , none was expected . the fit to a single isothermal model is only marginally acceptable , with @xmath16 for 32 degrees of freedom . residuals have a high - low - high pattern which signals the need for another smooth spectral component . when a second thermal component is added , best - fit parameters are @xmath17 kev , and @xmath18 kev , with the second component accounting for a minor fraction , roughly 1% of the 350 kev flux . for this fit , @xmath19 , lower by 7.5 than the value obtained in the single r - s model . the f - test probability of the second component is 0.94 for two additional degrees of freedom . although the @xmath20 is acceptable for this fit , the 90% error bounds of [ 16,80 ] kev for @xmath21 lie outside - in fact , well outside - the more tightly constrained values of @xmath22 kev obtained with _ asca _ ( markevitch 1996 ) , @xmath23 kev with _ rosat _ ( elbaz et al . 1994 ) , @xmath24 kev with _ xmm _ ( pratt 2001 ) , and @xmath25 kev measured with _ ( govoni et al . we therefore view this solution as doubtful . indeed , the range of solutions is very wide because the problem is numerically highly degenerate . if we consider the possible sets of ( @xmath21 , @xmath26 ) , the value @xmath27 defines the joint 90% probability contour in parameter space . the lowest temperature combination permitted in this range is @xmath28 kev and @xmath29 kev , with fractions of the 350 kev flux , respectively , of 99% and 1% . the highest temperature combination has @xmath30 kev and @xmath26 unbounded , with 13% of the 350 kev flux in the higher temperature component . these results seem to indicate that the two temperature fit improves on a single temperature one mostly by accounting for emission at the higher energy channels . when the second component is a power law the best - fit has @xmath31 , fully 11.0 lower than for the isothermal fit . the best - fit power law index is @xmath32 with 90% confidence bounds of [ -2.4 , 2.7 ] , and the 3 - 80 kev flux is @xmath33 erg/(@xmath34 s ) , with 90% confidence bounds of @xmath35 erg/(@xmath34 s ) . the f - test probability of this power law component is 0.99 for two additional degrees of freedom . the 350 kev flux of the power law component is 27% of the total . this component comprises most of the flux at energies @xmath36 kev . the iron abundance is @xmath37 , with 90% confidence bounds of @xmath38z_{\odot}$ ] . the best - fit fluxes from the combined thermal and power law emissions , and that of just the power law component , are shown in figure 1 , together with the data points . residuals of this best fit model are plotted in the lowest panel ; for comparison , the residuals to the single temperature model are shown in the middle panel . differences between the latter two plots are apparent mostly at high energies . the results of the temperature mapping with _ xmm _ ( pratt 2001 ) and _ chandra _ ( govoni 2004 ) clearly necessitate inclusion of a temperature distribution in the fits . the rxte fov includes a region which is much larger than that probed by both _ xmm _ and _ chandra_. this , and the fact that the rxte lacks spatial resolution , make the rxte insensitive to even large scale features in the _ chandra _ temperature map . the above fit to a two - temperature model grossly samples substantially different temperature components . just so as to verify that we can not get additional useful information from a more detailed spectral modeling , we have considered also a model with three components - two r - s and a power law . when the model parameters are fully unconstrained their values are obviously very loosely determined : the best - fit to this model - with @xmath39 for 28 dof - yields @xmath40 kev , a negligible second thermal component with @xmath41 kev , @xmath42 , @xmath43 , and power law normalization of @xmath44 erg/(@xmath34 s ) . the decrease in @xmath20 when a power law component is added is @xmath45 , so ( from f - test statistics ) the significance is about 85% ( or , more specifically , 83% - 86% , depending on the particular choice of ` interesting ' parameters ) . however , the @xmath20 hypersurface around the minimum correspoding to this parameter set is not very deep , meaning that a small increase of @xmath20 above its minimum value results in a rather different parameter set . in spite of the implied substantial uncertainty , we find that the need for a power law component is removed only when @xmath46 kev , which we consider to be rather unrealistic . a power law component at the above deduced level is apparent in the hexte measurements . to check general consistency between results from the pca and hexte , we have determined the parameters of this component by analyzing hexte data separately . clearly , due to the relatively high ( 12 kev ) lowest energy channel of hexte , an appreciable fraction of the thermal emission can not be detected by this experiment , so we have set the parameters of the isothermal component to its best - fit values from the joint ( pca & hexte ) data . fitting for the power law , we obtain @xmath47 , and @xmath48 erg/(@xmath49 s ) for the power law ( photon ) index and the integrated 3 - 80 kev flux . thus , hexte data alone provide significant additional evidence for nt emission whose spectral index has a narrower 90% confidence interval than that determined in the joint fits . while this is a constrained fit , with only two free parameters , it is both physically ( by virtue of the anticipated levels of thermal and nt emissions ) and practically ( given the relative insensitivity of hexte to emission at low energies ) well motivated . [ cols="<,^,^,^",options="header " , ] we note that a relatively short @xmath50 ks observation of a2163 with bepposax / pds yielded an upper limit on power law emission , @xmath51 erg/(@xmath49 s ) ( feretti 2001 ) . our much deeper observation yields a significant detection of nt emission with a flux level @xmath52 erg/(@xmath49 s ) , whose 90% confidence interval includes the pds upper limit . the results of the above spectral analysis clearly indicate that the emission in a2163 can not be described by a single temperature emission model ( as is apparent from positive residuals at both low and high energies ) . a better fit is provided by a two - temperature model , with the secondary thermal component accounting for the low energy residuals . temperature map of the central cluster region shows a region , roughly @xmath53 in area , where the temperature is @xmath54 kev . the secondary thermal component we deduce has a somewhat lower value , but given the large uncertainty in the value we deduce , the difference is not significant , as is the _ rough _ - due to the substantially larger ( @xmath55 fwhm ) rxte fov , which includes the emission from the cooler outer regions - consistency of the relative flux values of the two thermal components . however , in the best fitting model the second component is not thermal , but rather a nt power law . indeed , nt emission is expected from the populations of electrons that emit both the extended ic radio emission , as well as the emission from the several powerful radio sources in the central @xmath56 mpc region of a2163 . in table 2 we list the measured radio fluxes and spectral ( energy ) indices for the dominant sources in the rxte fov . in addition to the remarkably luminous emission from the most extended , regularly - shaped ` halo ' region , we include the emission from four sources measured by feretti et al . fluxes and spectral indices for the ` halo ' and relic sources are determined from vla measurements at 6 & 20 cm . to estimate the expected x - ray fluxes also from the three tailed ( t ) sources , we need to specify their spectral indices ; since these were not given by feretti , we adopt a typical value of @xmath57 . l@ c@ c@ c@ source & flux density - 20 cm ( mjy ) & spectral ( energy ) index & size ( kpc ) + halo & 1552 & 1.60.3 & 207070 + relic & 18.70.3 & 2.10.3 & 46040 + j1615 - 062 ( t1 ) & 34.50.5 & 0.80.1 & 450 + j1615 - 061 ( t2 ) & 6.00.3 & 0.80.1 & 120 + j1616 - 062 ( t3 ) & 24.90.4 & 0.80.1 & 270 + on the relevant spatial scales the radiation field most tightly coupled to the electrons by compton scattering is the cmb , coupling that is only further enhanced in a2163 ( @xmath4 ) due to the @xmath58 dependence of the cmb energy density . in order to quantify the emission that effectively results from the scattering , we need to determine the spectral density distribution of the electrons . radio emission from the central extended source dominates the overall emission , and since we expect the mean , volume - averaged field in this extended source to be much lower than field strengths in the galactic radio sources , it is clear that most of the compton - produced x - ray emission comes from the halo. nonetheless , we have estimated the emission by electrons in the other sources listed in table 2 . compton fluxes of the radio sources in table 2 can be estimated from their measured radio fluxes and estimated mean field values . the latter can be determined by assuming energy equipartition between particles and fields . this assumption may be roughly valid in the particles galactic sources , by virtue of the relatively short timescales of all the relevant processes governing the particles and fields - acceleration , couplings to the field and interstellar gas . ( we note in passing that the attainment of equipartition in the ic space is questionable : ic fields and particles are likely to be of galactic origin , but the evolution during their expulsion from galaxies and in ic space is quite different ; effective equilibration is not likely to occur there . ) using the deduced values of the equipartition in the radio sources we have estimated their respective compton x - ray fluxes . these turn out to be quite small for the relic as well as the three tailed sources , whose combined flux in the 3 - 80 kev band is @xmath59 of the measured value . ( note that consideration of the dominant non-halo radio sources in the clusters in which we have previously found evidence for nt emission - coma , a2256 , and a2319 - similarly shows that their predicted x - ray emission contributes negligibly to the deduced nt emission in each of these clusters . ) since the known radio sources in the central region of a2163 do not contribute appreciable x - ray emission , we attribute all the measured emission to electrons in the halo , and proceed to determine the mean , volume - averaged field in this region , @xmath60 , using the compton - synchrotron formulae ( rephaeli 1979 ) . for consistency with the basic premise that most of the measured compton flux is due to electrons in the ` halo ' , and the fact that the electron spectrum can be more precisely deduced from the radio data , we use the radio spectral index in order to infer the electron index . doing so we compute @xmath61 g . it should be emphasized that the overall uncertainty in the estimated mean field is substantially higher than this formal error due to the lack of spatial information on nt x - ray emission , which necessitates invoking the assumption that the emitting electrons and the field are co - spatial . as a consequence of the assumption that the spatial factors in the theoretical expressions for the radio and nt x - ray fluxes are roughly equal , it follows that the mean value of the deduced magnetic field is independent of the source size and distance . the electron energy density does depend on these parameters . scaling to the observed radius of the diffuse radio emission , and integrating the electron spectral distribution over all energies above 1 gev , we obtain @xmath62 , where @xmath63 is the value of the hubble constant in units of @xmath64 kms@xmath65mpc@xmath65 ) . based on the high galactic proton to electron energy density ratio of cosmic rays , it can be conjectured that the energetic proton energy density much higher than this value . a more specific assessment can not be made without consideration of the electron and proton origin and their respective energy losses . ic magnetic fields can also be estimated from faraday rotation measurements of background radio sources seen through clusters , yielding a different mean field value , @xmath66 . these measurements usually yield field values that are a few @xmath67 g ( see , , clarke , kronberg , and bhringer 2001 , and the review by carilli & taylor 2002 ) , up to an order of magnitude higher than values of @xmath60 . average field strength of a few @xmath67 g over an extended cluster region would have important implications on the range of electron energies that are deduced from radio measurements , and therefore on the electron ( synchrotron and compton ) energy loss times . weaker mean ( volume - averaged ) fields imply higher electron energies , and therefore shorter electron energy loss times , with possibly important ramifications for electron models ( , rephaeli 1979 , sarazin 1999 , ensslin 1999 , brunetti 2001 , petrosian 2001 ) . however , the apparent discrepancy between deduced values of @xmath60 and @xmath66 is usually naively interpreted . @xmath60 and @xmath66 actually provide very different measures of the mean field strength : @xmath60 is essentially a weighted volume - average of the electron density and of @xmath68 , with @xmath69 ; on the other hand , @xmath66 is an average of the product of the line of sight component of the field and the gas density . in general , the overall spatial dependence of these averages are considerably different , and so are their deduced values . specific examples and quantitative comparisons , including also the impact of observational uncertainties , are given by goldshmidt & rephaeli ( 1993 ) , and by newman , newman & rephaeli ( 2002 ) . [ brunetti , g. , 2001 , mn , 320 , 365 ] [ carilli , c.l . , & taylor , g.b . 2002 , araa , 40 , 319 ] [ clarke , t.e . , kronberg , p.p . , & bhringer , h. 2001 , apj , 547 , l111 ] [ elbaz d. , 1995 , a&a , 993 , 337 ] [ ensslin t.a . , 1999 , a&a , 344 , 409 ] [ ferreti l. , fusco - femiano g. , giovannini g. & govoni f. 2001 , 373 , 106 ] [ fusco - femiano , r. , 1999 , apj , 513 , l21 ] [ fusco - femiano , r. 2000 , apj , 534 , l7 ] [ fusco - femiano , r. 2003 , proceedings of ` matter and energy in clusters of galaxies ' , asp , conf . ser . , eds : s.bowyer & c-y.hwang ] [ fusco - femiano , r. , 2004 , apj , 602 , 73 ] [ goldshmidt , o. , & rephaeli , y. , 1993 , apj , 411 , 518 ] [ govoni , f. 2004 , apj , 605 , 695 ] [ gruber , d.e . 2001 , apj 562 , 499 . ] [ gruber , d.e . , & rephaeli , y. 2002 , apj , 565 , 877 ] [ herbig , t. , & birkinshaw , m. 1994 , baas , 185 , 530 ] [ kaastra , j.s . 1999 , apj , 519 , l119 ] [ macdonald , d. , & gruber , d.e . 2000 , baas , 32 , 1184 ] [ markevitch , m. 1996 , apj , 465 , l1 ] [ markevitch , m. and vikhlinin , a. 2001 , apj , 563 , 95 ] [ newman , w.i . , newman , a.l . , & rephaeli , y. 2002 , apj , 575 , 755 ] [ petrosian , v. 2001 , apj , 557 , 560 ] [ pratt , g.w . , arnaud , m. , & aghanim , n. 2001 , xxi moriond astrophysics meeting , eds . neumann & j.t.t . van ] [ rephaeli , y. 1977 , apj , 212 , 608 ] [ rephaeli , y. 1979 , apj , 227 , 364 ] [ rephaeli , y. , & gruber , d.e . 2002 , apj , 579 , 587 ] [ rephaeli , y. , & gruber , d.e . 2003 , apj , 595 , 137 ] [ rephaeli , y. , gruber , d.e . , & blanco , p.r . 1999 , apj , 511 , l21 ] [ rossetti , m. , and molendi , s. 2004 , a&a , 414 , l41 ] [ sarazin , c.l . 1999 , apj , 520 , 529 ]
a2163 was observed by the rxte satellite for @xmath0530 ks during a 6 month period starting in august 2004 . the cluster primary emission is from very hot intracluster gas with @xmath1 kev , but this component does not by itself provide the best fitting model . a secondary emission component is quite clearly needed , and while this could also be thermal at a temperature significantly lower than @xmath1 kev , the best fit ( to the combined pca and hexte datasets ) is obtained with a power law secondary spectral component . the deduced parameters of the non - thermal ( nt ) emission imply a significant fractional flux amounting to @xmath2 of the integrated 3 - 50 kev emission . nt emission is expected given the intense level of radio emission , most prominently from a large extended ( ` halo ' ) central region of the cluster . interpreting the deduced nt emission as compton scattering of the radio - emitting relativistic electrons by the cmb , we estimate the volume - averaged value of the magnetic field in the extended radio region to be @xmath3 g . _a&a _
symmetry classification of possible gap structures established the framework for separating conventional superconductors from their unconventional counterparts @xcite . superconductors with the order parameters that transform according to the trivial irreducible representation of the point symmetry group of the crystal are usually labeled conventional . if the order parameter transforms according to a non - trivial representation of the same group , the superconductor is called unconventional . in the latter class the symmetry of the order parameter is lower than the full group symmetry . this lowering of symmetry is usually believed to be related to strong repulsive coulomb interactions and specific pairing mechanisms ( for example , due to magnetic fluctuations ) . as a result , determining the gap structure is one of the crucial steps in testing our understanding of the origin of superconductivity in a given material . an important group of unconventional superconductors are those with order parameter that vanishes ( has nodes ) for some directions at the fermi surface . thermal conductivity is an exceptionally powerful probe for testing the shape of the gap in nodal superconductors since only the unpaired electrons ( with momenta close to the nodal directions ) carry entropy . the presence or absence of the nodes , and sometimes their type ( point or line ) , can be inferred from the dependence of the thermal conductivity on the magnetic field and temperature . an additional test for the existence of line nodes is the so - called universality of the low - temperature coefficient of the thermal conductivity , @xmath2 @xcite . the universal behavior refers to the relative insensitivity of @xmath3 to the concentration of the impurity atoms and to the details of the scattering on individual impurity centers @xcite . physically , since impurity scattering is pairbreaking , it generates near - nodal quasiparticles which can carry heat , and whose lifetime is , in turn , is limited by scattering on the same impurities . for linear nodes the two effects cancel nearly exactly . in the high - t@xmath4 cuprates , for example , the quantitative agreement between theoretical estimates of @xmath3 and the measured thermal conductivity was established @xcite . however , determining the location of the nodes in the momentum space from such measurements is much harder . the most direct tests measure the variation in the heat transport with the orientation of the applied magnetic field with respect to the nodal directions @xcite . long before such experiments were attempted , it was proposed that the _ anisotropy _ of the thermal conductivity along two different directions as a function of temperature allows to infer information about the gap structure @xcite . essentially , the measurement determines the predominant direction of the fermi velocity for the nodal quasiparticles . combined with the knowledge of the fermi surface , measurements of the evolution of the anisotropy in the superconducting state impose stringent constraints on the possible loci of the nodes . this method was applied to upt@xmath5@xcite , and the results argued convincingly for a line of nodes at @xmath6 . the main reason this experiment did not uniquely determined the gap structure is that theoretically expected results depend on the detailed shape of the fermi surface and the superconducting gap , i.e. on the basis functions for the particular representation @xcite . this detail apart , it is believed that the measurement provides a good test for the horizontal vs. vertical linear nodes . very recently the anisotropy of the thermal conductivity along inequivalent crystalline directions was measured in heavy fermion ceirin@xmath0 by shakeripour et al . the ratio of the @xmath7-axis to the in - plane thermal conductivity , @xmath8 , rapidly decreases in the superconducting state . similarity of the anisotropy evolution to that for upt@xmath5 , where equatorial line of nodes is believed to exist , led the authors of ref . to suggest that the superconducting gap in ceirin@xmath0 also has a horizontal line of nodes . this is surprising since cecoin@xmath0 , a close relative of ceirin@xmath0 , has vertical lines of nodes @xcite , and the prevailing belief is that the two compounds are quite similar ( although some evidence points to different origin of superconductivity in the two systems @xcite ) . consequently , in this paper we re - visit the analysis of the temperature dependence and the anisotropy of the thermal conductivity in nodal superconductors . since both de haas - van alphen measurements , and band structure calculations show that ceirin@xmath0 has an open , quasi - two dimensional piece of the fermi surface with significant @xmath9-electron contribution @xcite , our main focus here is on such anisotropic systems . we discuss the implications of the measurements for the symmetry of the order parameter , and compare the anisotropy of the thermal conductivity for several models relevant not only to ceirin@xmath0 , but also to other systems with a quasi - two dimensional fermi surface . the remainder of the paper is organized as follows . in next section we introduce the basic experimental facts and theoretical considerations , formulating the models . the subsequent sections are devoted to the analysis of the proposed models . we then critically compare our results with experiment , and propose further measurements to test the gap symmetry in ceirin@xmath0 . ceirin@xmath0 is a member of the so called 115 series , which also include cecoin@xmath0 and cerhin@xmath0 . while in the rh system the @xmath9-electrons of ce remain localized and undergo antiferromagnetic ordering , both ir and co compounds are paramagnetic heavy fermion metals . in both systems de haas - van alphen measurements and the band structure calculations indicate that a major sheet of the fermi surface is quasi - two dimensional , although the energy dispersion along the @xmath7-axis is substantial , as evidenced by a very moderate anisotropy ( @xmath10 ) between the out of plane and in - plane normal state transport coefficients . in contrast to cecoin@xmath0 , which is a highly unconventional metal , likely close to a quantum critical point at finite magnetic field @xcite , ceirin@xmath0 near the superconducting phase is a good fermi liquid , and therefore its excitations should be adequately described by the calculations in the fermi - liquid framework . both ceirin@xmath0 and cecoin@xmath0 are ambient pressure superconductors , and knight shift measurements indicate singlet pairing @xcite . specific heat and in - plane thermal conductivity measurements @xcite , as well as penetration depth @xcite and spin - lattice relaxation rate @xcite , show the existence of line nodes . in cecoin@xmath0 there is overwhelming evidence for the vertical line nodes from the anisotropy of the thermal conductivity and specific heat under rotated magnetic field @xcite , and from the tunneling into the andreev bound states @xcite . given the strong similarities between the fermi surfaces it is tempting to conclude that the gap structure is similar in the two materials . there is , however , evidence pointing towards differences in the origin ( and hence possibly in the type ) of the superconducting state of the co and ir compounds @xcite . authors of ref . measured the temperature dependence of the thermal conductivity in ceirin@xmath0 along two inequivalent crystalline directions , @xmath11 ( out of plane ) and @xmath12 ( in plane ) , and made two significant observations : a ) the ratio @xmath13 is nearly temperature - independent above the superconducting transition temperature , @xmath14 , but is rapidly reduced with @xmath1 at @xmath15 ( note that inelastic scattering yields a peak in thermal conductivity just below @xmath16 k , which may lead to the decrease of @xmath17 appearing very pronounced ) ; b ) the low @xmath1 limit of the in - plane @xmath18 appears universal , but the interplane @xmath19 does not show the universal limit . our goal below is to examine the possible origin and implications of this result . independently of the choice of theoretical model , the results a ) and b ) above imply , prima facie , that the effect of superconductivity on the quasiparticles with the fermi velocity predominantly in the plane is different from that for the quasiparticles moving along the c - axis : the states that carry entropy along the @xmath7-axis have a larger `` effective gap '' than those carrying heat current in the plane . several processes of different physical origin may lead to this behavior . in the following we do not consider inelastic scattering ; this is appropriate at low temperatures , @xmath20 . the range of temperatures in ref . where the inelastic scattering can be neglected is not immediately clear from the data , note , however , that , if the inelastic scattering is isotropic , our conclusions regarding the anisotropy of the thermal transport are not affected . since the fermi surface of ceirin@xmath0 has several sheets , it is , of course , possible that the multiband effects are responsible for the observed behavior . however , without a microscopic theory detailing the gaps on each of the sheets , reaching reliable conclusions about the agreement between theory and experiment seems impossible , while phenomenological multiband theory has too many fitting parameters to seriously constrain possible gap structures . we aim to construct a minimal model that captures the essential physics of the the experimental observations , and therefore restrict ourselves to considering one electronic band . since most band structure and dhva analyses suggest an important role of the quasi - two dimensional ( open along the @xmath7-axis ) band , we choose such a band for our approach . we approximate it by a simple nearest neighbor tight binding expression @xmath21 where @xmath22 are the components of the quasiparticle momentum , @xmath23 , @xmath7 is the lattice spacing in the @xmath24 direction , @xmath25 is the effective mass , and @xmath26 is the interplane tunneling matrix element . in all but one of our models we take @xmath26 to be momentum - independent . we show below however , that the model with @xmath27-dependent interplane tunneling that provides a viable path towards explaining the experimental results , even with the assumption of vertical , rather than horizontal , nodes . we consider singlet order parameters , and begin with the simple model of vertical line nodes similar to the order parameter in cecoin@xmath0 , introducing the azimuthal in - plane angle @xmath28 , * * model a : * @xmath29 , and + @xmath30 . we show in the next section that this model is incompatible with the experimental observations . physically , the fermi surface is cylindrically symmetric and the gap depends only on the azimuthal angle , @xmath31 , while the @xmath24-axis component of the fermi velocity is @xmath31-independent . as we show below in sec . [ sec : a ] , these conditions ensure that the ratio @xmath32 , is temperature - independent even in the superconducting state . this constraint needs to be relaxed to explain the experimental results , and we consider several possible models . first , we show that for a tight - binding fermi surface it can be expected , from microscopic considerations , that the gap also acquires a weak modulation along the @xmath7-direction , while retaining vertical lines of nodes , and therefore consider * * model b : * @xmath29 , and + @xmath33 . while in this model there exists a temperature - dependent anisotropy , we find that this dependence is generally weak , and hence unlikely to provide a satisfactory explanation of the experimental observations ( although we can not exclude it ) . in the second approach to lifting the constraints of model a we consider a situation with vertical line nodes , but where the nodal quasiparticles are less efficient than the normal electrons ( on average ) in transporting heat along the @xmath7-axis compared to the @xmath34 plane . to model this we note that the measured and calculated fermi surface of ceirin@xmath0 is not rotationally symmetric @xcite , and that the energy dispersion of the electrons along the @xmath24-axis clearly depends on the direction of the in - plane momentum . this dependence is expected in the @xmath35 symmetry , and @xmath26 is determined by the structure of the overlapping wave functions for the quasi - two dimensional band . we therefore introduce the angle dependence into the interplane tunneling , preserving the crystal symmetry in the plane , and propose * * model c : * @xmath36 , and + @xmath30 . for @xmath37 the interplane tunneling is smaller for the nodal quasiparticles than for antinodal ones , which ensures the @xmath1-dependence of the anisotropy ratio . the situation is somewhat reminiscent of that in the high - t@xmath4 cuprates where , in the absence of orthorhombic distortion , the interplane tunneling of the nodal quasiparticles is suppressed , i.e. @xmath38 @xcite , and therefore the observed temperature dependence of the josephson plasma resonance frequency @xcite and the @xmath7-axis penetration depth @xcite is different from that suggested by the simple density of states power counting . we show that such a model gives a significant temperature - dependent anisotropy of the thermal conductivity . finally , we consider a model with horizontal , rather than vertical line nodes , similar to the hybrid @xmath39 gap proposed for upt@xmath5 , with broken time - reversal symmetry . the basis function of the representation is described as @xmath40 . for open fermi surface we require periodicity along the @xmath24-direction , and the appropriate basis function is @xmath41 . the excitation spectrum and the thermal conductivity are only sensitive to the gap amplitude , @xmath42 , where @xmath43 . since our consideration of model b shows ( see the analysis below ) that weak modulation does not change the anisotropy qualitatively , we ignore it , and consider * * model d : * @xmath29 , and + @xmath44 . we show that , this model also gives a substantially temperature - dependent ratio @xmath45 . moreover , we point out that from the data of ref . it is impossible to distinguish model c from model d , and suggest further measurements to probe the gap structure in ceirin@xmath0 . we are now ready to consider each model in detail . in order to make connection with experiment , we largely focus on the thermal conductivity and its anisotropy @xmath46 we begin by briefly looking at the normal state properties . in ref . .,title="fig:",width=100 ] in ref . .,title="fig:",width=115 ] the fermi surfaces for the models a through d are shown in fig . [ fig : fs ] . in each case we transform to the cylindrical coordinates and parameterize the fermi surface by the azimuthal angle @xmath47 , and the c - axis quasimomentum @xmath48 . the jacobian of the transformation from variables @xmath49 to variables @xmath50 is unity , and therefore no angle - dependence of the normal state density of states appears . we denote the fermi energy @xmath51 , and define the fermi momentum at @xmath52 as @xmath53 . in models a , b , d above the tunneling @xmath54 is @xmath31-independent . in that case , the radius of the fermi surface is @xmath55 , and the components of the fermi velocity are @xmath56 [ vf ] where @xmath57 . in the normal state the anisotropy of the transport coefficients is given simply by @xmath58 where the average over the fermi surface is @xmath59 for the simple model above @xmath60 . we can obtain a very rough estimate of the parameter values relevant for ceirin@xmath0 from the dhva measurements and the band structure calculations @xcite . for the quasi - cylindrical piece of the fermi surface the typical radius is 1/5 of the brillouin zone size , @xmath61 , where @xmath62 is the in - plane lattice constant . in this material @xmath63 , so that @xmath64 . the normal state resistivity ratio , @xmath65 , implying @xmath66 . for model c the situation is slightly more complex . we still have @xmath67 and therefore @xmath68.\ ] ] at arbitrary in - plane @xmath69 the @xmath70-component of the velocity is @xmath71 and , consequently , at the fermi surface , @xmath72 since we work in the regime @xmath73 , we can ignore the corrections from the term @xmath74 in the denominator as they appear only at the order @xmath75 , and write @xmath76.\ ] ] for the conductivity anisotropy of 2.5 cited above , the _ maximal _ value of @xmath77 , when there is no @xmath78 component , is approximately @xmath79 , which corresponds to the contribution of the second term of less than 5% to the in - plane transport in the normal state . its effect in the superconducting state is even smaller , since the @xmath80-dependent part of @xmath81 vanishes along the nodal directions . consequently , we ignore the @xmath80-dependent contribution to the in - plane transport hereafter . the thermal conductivity of a superconductor with the gap function whose average over the fermi surface vanishes is given by @xcite @xmath82 here @xmath83 is the normal state density of states , and the renormalized frequency @xmath84 , where @xmath85 is the self - energy due to impurity scattering . for the model with @xmath86 , and the hopping that is @xmath31-independent , we immediately find for different orientations of the heat current @xmath87 \ ; , \end{aligned}\ ] ] where @xmath88 and @xmath89 are the complete elliptic integrals , and @xmath90 therefore the anisotropy of the thermal conductivity in the superconducting state remains unchanged compared to the normal state , @xmath91 . moreover , in this model the universal ( nearly independent on impurity concentration ) low temperature limit , obtained from eq.([eq : athermal ] ) , is @xcite @xmath92 for both in - plane and out of plane directions . in eq.([eq : kappauniv ] ) @xmath93 is the low - energy scattering rate , @xmath94 , and @xmath88 is the complete elliptic integral . for clean superconductors , @xmath95 , and we have @xmath96 so that @xmath97 , universal and independent of the impurity scattering rate . this constant and temperature independent ratio @xmath98 is in contradiction to the experimental measurements of ref . , and therefore model a does not provide a satisfactory description of the superconducting state of ceirin@xmath0 . the symmetry classification does not uniquely determine the functional form of the superconducting gap around the fermi surface . the gap can be given by any combination of the basis functions transforming according to the chosen irreducible representation . for example , in a cubic system , both a constant gap and that varying as @xmath99 have the same symmetry properties . of course , any @xmath27-space variation of the gap beyond that imposed by symmetry requirements costs condensation energy , and therefore , all other things being equal , the simplest basis functions often provide the most stable superconducting state . a conventional superconductor with a spherical fermi surface strongly prefers a fully isotropic gap , while a strongly anisotropic quasi two - dimensional system with a circular fermi surface and the dominant pairing in the @xmath100 channel is likely to have the gap with the @xmath101 momentum dependence . this is the reason for using the familiar notation of @xmath102-wave for the former case , and @xmath103 for the latter . on the other hand , superconductivity is an instability of the fermi surface towards pairing , and therefore it is natural to expect that in real materials with complex fermi surfaces the gap structure is also more complex . the correct gap can only be determined from the microscopic theory of superconductivity . to solve eliashberg equations in superconductors with electron - phonon mediated pairing , allen introduced the fermi surface harmonics @xcite , which have the symmetry of the fermi surface , are orthonormal , and are convenient basis functions for the expansion of the superconducting pairing field . in the absence of microscopic theory we use similar symmetry arguments , and go beyond lowest order basis functions . for simplicity , we continue to consider the @xmath103 gap symmetry , @xmath104 ; all our considerations apply equally well to other gaps with vertical line nodes , such as @xmath105 . choice of this symmetry determines the irreducible representation of the crystal point group . in a system with the fermi surface open along the c - axis , we need to add to this point group the invariance with respect to translations by a reciprocal lattice vector along @xmath24 . in the even ( odd ) channel the basis functions for translation are @xmath106 ( @xmath107 ) , with @xmath108 integer . therefore quite generally the singlet @xmath103 gap has the form @xmath109 similar interaction for @xmath102-wave pairing was considered by bulaevskii and zyskin @xcite , and for @xmath110-wave symmetry by rajagopal and jha @xcite . note that since even powers of @xmath48 , and therefore @xmath111 , transform according to a trivial representation of the tetragonal point group , the functions with different @xmath108 belong to the same representation and can be mixed . the degree of mixing is , of course , determined by the pairing interaction . we consider a separable model , @xmath112 where @xmath113 to illustrate the behavior of the model , and to be consistent with keeping only the nearest neighbor layer hopping in the energy dispersion , eq.([eq : energy ] ) , we truncate the expansion at @xmath114 , so that @xmath115\cos 2\phi\cos 2\phi^\prime . \label{eq : vmod}\ ] ] to the same accuracy the gap is given by @xmath116\cos 2\phi , \label{gapc - ab}\ ] ] and is determined from the self - consistency equation ( for pure superconductor in the weak - coupling limit ) @xmath117 here @xmath118 , @xmath119 are the fermionic matsubara frequencies , and @xmath120 is the cutoff energy for the pairing interaction . formally , the gap given by eq.([gapc - ab ] ) allows for a horizontal line of nodes when @xmath121 . this is , however , physically unlikely . in the interaction @xmath122 above , @xmath123 is the relative strength of the pairing in neighboring planes compared to the in - plane pairing , so that @xmath124 . generally we also expect @xmath125 @xcite , so that the gap modulation along the @xmath7-axis is insufficient to produce horizontal lines of nodes . inversely , if the interaction strengths in different channels in eq.([eq : vmod ] ) are comparable , we can not truncate the expansion at the lowest order terms . therefore , while in principle the symmetry considerations allow the gap in eq.([gapc - ab ] ) to have both horizontal and vertical nodes , the model considered here can only be used if it yields small to moderate values of @xmath126 . consider first the linearized equations for the transition temperature , @xmath14 . introducing dimensionless coupling constant @xmath127 , we find that @xmath14 and @xmath126 satisfy @xmath128 \ln\frac{2\gamma_e\omega_0}{\pi t_c } , \\ \delta&=&\frac{g}{2}\left[\lambda_1+\frac{\lambda_2\delta}{2}\right ] \ln\frac{2\gamma_e\omega_0}{\pi t_c}.\end{aligned}\ ] ] [ linearizedeq ] here @xmath129 is the euler s constant . note that in the absence of the @xmath130 term in the pairing interaction the equations for the temperatures of the transition into the states @xmath131 and @xmath132 decouple , and each is reduced to the standard bcs expression , @xmath133 , with the effective coupling constants in the @xmath134 channels @xmath135 and @xmath136 for @xmath137 and @xmath114 respectively . for @xmath138 this implies that @xmath139 , so that the simple @xmath140 gives the dominant order parameter . for the generic case @xmath141 , however , the two channels are coupled , and the transition occurs directly into the c - axis modulated state @xmath142 . the transition temperature is @xmath143 where @xmath144 we assumed that the coupling @xmath123 is not too repulsive . the transition temperature depends on the magnitude , but not on the sign of @xmath130 . on the other hand , the modulation , @xmath145 , depends on the sign of @xmath130 , and can be either positive or negative . the modulation amplitude , @xmath126 , is generally @xmath1-dependent . consider the self - consistency equation at @xmath146 , @xmath147 \frac{\left[1+\delta\cos\chi_z^\prime\right]\cos^2 2\phi^\prime } { \sqrt{\omega_n^2+\delta_0 ^ 2(1+\delta\cos\chi_z^\prime)^2\cos^2 2\phi^\prime } } \\ \delta = g\pi t\sum_{|\omega_n|<\omega_0 } \int_{-\pi}^\pi \frac{d\chi_z^\prime}{2\pi } \int_0^{2\pi}\frac{d\phi^\prime}{2\pi } \left[\lambda_1 + \lambda_2 \cos\chi_z^\prime\right ] \frac{\left[1+\delta\cos\chi_z^\prime\right]\cos^2 2\phi^\prime } { \sqrt{\omega_n^2+\delta_0 ^ 2(1+\delta\cos\chi_z^\prime)^2\cos^2 2\phi^\prime } } \end{aligned}\ ] ] [ self - cons1 ] in the linearized form considered above , all the terms in the integrand linear in @xmath148 , vanish by symmetry . on the other hand , below @xmath14 , when @xmath149 term in the denominator is important , these terms give a finite contribution to the gap equation . moreover , in contrast to the conventional `` one - channel '' systems , the cutoff frequency @xmath120 can not be removed from the self - consistency equations completely . as a consequence , @xmath126 acquires temperature dependence . of course , in the weak coupling limit this dependence is insignificant : in both integrals in eq.([self - cons1 ] ) the greatest contribution is from the range @xmath150 , and therefore essentially insensitive to the absence or presence of the gap . in most situations the dependence on the cutoff vanishes already for @xmath151 . this is seen from the self - consistently determined temperature evolution of the gaps in the two channels computed using eqs.([self - cons1 ] ) and shown in the left panel of fig.[fig : del12 ] . the corresponding plot of the modulation , @xmath126 , in the right panel of fig.[fig : del12 ] demonstrates that the modulation is temperature - independent for most parameter values . we find that one exception is the somewhat artificial situation when a ) the bare transition temperatures into the two states are very close , @xmath152 so that @xmath153 , _ and _ b ) @xmath154 . while @xmath126 is still temperature - independent for @xmath155 , the temperature variations persist to large values of the cutoff , see fig.[fig : del12 ] . part of the reason is that the value of @xmath156 at @xmath14 obtained above depends on the order in which the limits are taken : setting @xmath157 first gives @xmath158 , while taking the limit @xmath159 first for a finite value of @xmath160 yields @xmath161 . physically , at the point @xmath162 the linearized equations , eq.([linearizedeq ] ) , decouple , and give no information on the values of @xmath163 and @xmath164 . as is well known , the gap amplitudes are determined by the fourth order terms in the ginzburg - landau expansion of the free energy , or , equivalently , by the third order terms in the gap equations . we find that inclusion of such terms gives three different possible solutions for @xmath163 and @xmath165 . two of them are trivial , @xmath166 ( lowest energy , as expected ) , and @xmath167 . however , we find that there is a third solution with @xmath168 , which has the highest energy of the three . at the same time for @xmath157 and any @xmath130 @xmath169 corresponds to the lowest free energy , which suggests a singular limit @xmath159 . this is confirmed by carrying our the expansion of @xmath126 in @xmath170 , at finite @xmath130 , where we find that the coefficients are singular for @xmath157 . this suggests a strong dependence of the energies of the three stationary points on parameters , and on temperature . : for different cutoffs @xmath171 the curves lie on top of one another . right panel : temperature dependence of the gap anisotropy , @xmath126 for different coupling constants . note the temperature variation of the anisotropy in the regime of moderately large @xmath172 and small @xmath173 , where we present the results for @xmath171 . , width=312 ] away from the non - physical point @xmath162 , the numerical solution of the self - consistency equations , eqs.([self - cons1 ] ) shown in fig.[fig : del12 ] makes it clear that we can use the gap @xmath174\cos 2\phi$ ] with temperature independent @xmath126 . however , at low temperatures @xmath175 is no longer at the bcs ratio of 2.14 , but depends on the value of @xmath126 . this dependence is computed numerically , shown in fig.[fig : delta - delta ] , and is used in subsequent computation of the thermal conductivity . . note the smooth evolution of the gap amplitude across the value @xmath176 , where a horizontal line of nodes appears.,width=288 ] we use eq.([eq : genthermal ] ) with @xmath174\cos 2\phi$ ] to evaluate the temperature dependence of the thermal conductivity in a quasi - two dimensional superconductor . for our calculations we used the impurity scattering in the unitarity limit , with the normal state @xmath177 . while both @xmath178 and @xmath179 show the standard dependence on the scattering rate for unconventional superconductors @xcite , the anisotropy ratio is essentially insensitive to impurity concentration in the clean limit @xmath180 . the results are shown in figs . [ fig : bkapx ] and [ fig : bkapz ] . one point of note is that the @xmath7-axis thermal conductivity is even in @xmath126 , while the in - plane transport is sensitive to whether the largest or the smallest gap occurs for @xmath6 , where the in - plane fermi velocity is the greatest , see eq.([vf ] ) . of immediate interest to us is the temperature dependence of the anisotropy , @xmath181 , which is shown in fig . [ fig : banis ] . the qualitative understanding of the temperature dependence relies on the observation that the modulation changes the gap most significantly in the regions @xmath182 , where the quasiparticle velocity is strictly in the ab plane , with no component along the @xmath7-axis . consequently , for @xmath183 the in - plane conductivity is reduced compared to that of a superconductor with an unmodulated gap , while for @xmath184 it is increased . as a result , for @xmath183 the anisotropy ratio @xmath45 below @xmath14 is enhanced compared to the normal state value , while for @xmath184 it is reduced . notice that for our model @xmath178 is insensitive to the sign of @xmath126 since points at the fermi surface where @xmath185 ( @xmath186 ) have identical values of the fermi velocity , and therefore contribute equally to the @xmath7-axis thermal conductivity . the residual anisotropy in the @xmath187 limit can be computed analytically once we recognize that , prior to integration over the @xmath24-component of the momentum , for each value of @xmath188 the contribution of the kernel to the conductivity is universal in complete analogy to a two - dimensional @xmath110-wave superconductor with the gap @xmath189 . consequently , the ratio of the residual low temperature terms is ( @xmath190 ) @xmath191/ \left[\int_{-\pi}^{\pi } \frac{(1+\lambda\cos ( \chi_z))d\chi_z}{1+\delta\cos\chi_z}\right ] \\ & = & 2\frac{1-\sqrt{1-\delta^2}}{\delta^2 } \left[\frac{\lambda}{\delta}+\left(1-\frac{\lambda}{\delta}\right)\frac{1}{\sqrt{1-\delta^2 } } \right]^{-1}.\end{aligned}\ ] ] this result is shown in fig . [ fig : banis ] . the experimentally determined @xmath192 is as low as 0.5 at temperatures of the order of 0.2t@xmath4 @xcite . this can only be achieved for large negative values of the gap anisotropy , @xmath193 , when the gap nearly vanishes in the equatorial plane of the fermi surface . this value is possible for sufficiently strong coupling @xmath130 , but we consider it not very likely in ceirin@xmath0 . in this model both @xmath178 and @xmath179 are universal , albeit with different values , and therefore future doping studies of this material will be able to test the universality of the @xmath24-axis thermal conductivity . in quasi - two dimensional systems the interplane matrix element depends on the overlap of the atomic orbitals contributing to the bands close to the fermi surface . therefore , in general , the in - plane and out - of - plane motion of the quasiparticles are coupled , and , in writing the tight - binding ansatz for the open fermi surface along the c - axis , it is necessary to recognize that the matrix element @xmath54 is a function of the in - plane direction of the momentum , @xmath31 . moreover , simply looking at the quasi-2d sheet of the fermi surface of ceirin@xmath0 as obtained by band - structure calculations and by dhva measurements @xcite , it is clear that the fermi surface is not rotationally invariant , and therefore it is necessary to account for the angle - dependence of tunneling . for the fermi surface of a general shape @xmath194 the difference between the maximal and minimal values of the radius , @xmath195 is @xmath196 . in the 115 series this difference is greater along the [ 110 ] direction than along the [ 100 ] direction , which suggest that @xmath54 is @xmath31-dependent @xcite . in the normal state this dependence does little beyond replacing @xmath54 with the appropriate average over the directions , see sec . [ sec : normal ] . in a superconducting state with an anisotropic gap , however , the effect of the modulation of the interplane hopping can be much more pronounced . the best studied example is the high - t@xmath4 cuprates , where , on the basis of band structure calculations , @xmath197 is expected to vanish along the directions of the nodes in the gap @xcite . among the manifestations of this effect are the high power law in the temperature variation of the c - axis penetration depth ( @xmath198 rather than @xmath1 in the pure limit ) @xcite , and weak dependence of the josephson plasma resonance frequency on temperature at @xmath20 @xcite . in the analysis of the thermal conductivity it is easy to see that any directional dependence @xmath197 results in a deviation of the anisotropy ratio , @xmath98 , from its normal state value , @xmath199 , even for a superconducting gap that has pure @xmath110-wave form , @xmath200 , with no @xmath48-dependence . this is simply because the quasiparticle velocity , @xmath201 , now acquires a dependence on the angle @xmath31 , which affects the evaluation of the thermal conductivity kernel , @xmath202 , see eq.([eq : genthermal ] ) . moreover , since the universal low temperature behavior is due to near - nodal quasiparticles , any suppression of the hopping matrix element in the vicinity of the nodes reduces the contribution of these quasiparticles to the c - axis transport , making it non - universal , while preserving the universality of the low temperature limit for the in - plane conductivity . therefore assuming a modulated hopping may provide a route towards the explanation of the experimental results . once again we adopt the model energy dispersion of the form @xmath203 where @xmath204 which is the simplest anisotropic form satisfying the tetragonal symmetry . this additional @xmath31 dependence does not affect the in - plane thermal conductivity , @xmath179 , but does modify the out of plane @xmath178 . as discussed in sec.[sec : normal ] , to a good approximation the normal state anisotropy ratio is @xmath205 ( k_f c)^2.\ ] ] for simplicity , and in agreement with the results of the previous section showing that small @xmath7-axis modulation of the gap does not yield significant corrections to the anisotropy ratio in the superconducting state , we consider a @xmath110-wave gap in this calculation . in agreement with the discussion above , we are interested in the situation when the interplane transport is suppressed along the nodal directions , and therefore take @xmath200 . neglecting the small corrections to the in - plane fermi velocity due to the angle - dependence of @xmath54 , as discussed in sec.[sec : normal ] , we find @xmath206 \ ; .\ ] ] the kernel for the interplane conductivity is more complex , @xmath207 % \\ + 2t_0t_1\left[\tiw \ : i_{1e } \left ( \frac{\delta_0}{\tiw } \right ) + \frac{|\tiw|^2 - \tiw^2}{2\tiw } \ : i_{1k } \left ( \frac{\delta_0}{\tiw } \right)\right ] \\ + t_1 ^ 2\left[\tiw \ : i_{2e } \left ( \frac{\delta_0}{\tiw } \right ) + \frac{|\tiw|^2 - \tiw^2}{2\tiw } \ : i_{2k } \left ( \frac{\delta_0}{\tiw } \right)\right ] \biggr\},\end{aligned}\ ] ] where @xmath208 the @xmath209 limit , which is important for the comparison with the universal limit , can also be readily evaluated as a function of the low energy scattering rate , @xmath93 , by setting @xmath210 . we reproduce the standard result for the in - plane conductivity , @xmath211 while for the interplane thermal conductivity we find @xmath212 \\ & & + t_1 ^ 2\frac{\gamma^2}{\delta_0 ^ 2 } \left [ e\left(\frac{\delta_0}{\sqrt{\gamma^2+\delta_0 ^ 2}}\right ) -2\frac{\gamma^2}{\delta_0 ^ 2}\left [ k\left(\frac{\delta_0}{\sqrt{\gamma^2+\delta_0 ^ 2}}\right ) -e\left(\frac{\delta_0}{\sqrt{\gamma^2+\delta_0 ^ 2}}\right ) \right ] \right ] \biggr\}.\end{aligned}\ ] ] the normal state scattering rate.,title="fig:",height=240 ] [ fig : ckap ] in the clean limit , @xmath213 , the residual in - plane conductivity is universal , @xmath214 while the @xmath7-axis conductivity , for @xmath215 , is given by @xmath216 which implies that the residual anisotropy ratio is @xmath217^{-1}<1.\ ] ] even for a very moderate anisotropy ratio , @xmath218 ( which , for samples with @xmath219 is well within the range of the approximation ) , the anisotropy ratio at low temperature is already low at about 22% . the reduction of the anisotropy is even more pronounced for the cases of extreme anisotropy , @xmath220 , when @xmath221\ll 1 . \label{ranis : c}\ ] ] consequently in this model the residual anisotropy ratio can be arbitrarily small depending on the purity . if there are no symmetry requirements for the hopping matrix element to vanish for the nodal directions , in most situations we do not expect @xmath78 and @xmath80 to differ by an order of magnitude ; this suggests that the residual linear term is visible in the @xmath7-axis thermal conductivity , but may be significantly smaller than its in - plane counterpart . while the results above give a rough estimate of the effect of impurities on the residual anisotropy , it is worth noting that disorder is likely to make the interplane hopping more isotropic , and therefore tends to restore the anisotropy to the the normal state value faster that what eq.([ranis : c ] ) indicates . .,title="fig:",height=240 ] [ fig : crs ] so far we looked at the models with vertical line nodes , motivated by possible similarities between ceirin@xmath0 and cecoin@xmath0 , and attempted to reconcile them with the experimental measurements . of course , in the absence of any information about the nodal structure , a natural explanation for the observed anisotropy is that , in analogy to upt@xmath5 , the system has a horizontal line of nodes , and therefore the nodal quasiparticles do not contribute to the @xmath7-axis transport . we consider this model now . the hybrid gap belongs to the representation that transforms as @xmath222 . the basis function for this representation over an open fermi surface is @xmath223 , and therefore , for , if we take the weakly modulated limit , @xmath224const , the gap amplitude is @xmath225 . if we take into account the fourfold , rather than cylindrical , shape of the fermi surface , there may be additional small modulation of this gap with the component of the fermi momentum in the @xmath34 plane as a function of @xmath24 : based on our results for model b in sec.[sec : b ] , we ignore these . then the density of states , and the in - plane thermal conductivity for this model are identical to that for a system with vertical line nodes , while the interplane thermal conductivity kernel is given by @xmath226\ ] ] the temperature dependence of @xmath179 and @xmath178 is shown in fig . [ fig : dkap ] , and the corresponding anisotropy ratio for the hybrid gap is given in fig . [ fig : drs ] for the hybrid gap for different values of the impurity scattering . at low @xmath1 the in - plane thermal conductivity is universal , while the interplane conductivity , @xmath227 \\ & \approx & \frac{2\pi}{3}\frac{n_0t^2c^2}{\delta_0}\frac{\gamma^2}{\delta_0 ^ 2 } \ln\frac{4\delta_0}{e\gamma},\end{aligned}\ ] ] gives the residual anisotropy @xmath228 this result is clearly non - universal . notice that the dependence on the impurity scattering here differs from the @xmath229 result obtained for upt@xmath5 in ref . . the reason for the discrepancy is that for the hybrid gap and a fermi surface closed along the @xmath7-axis the quasiparticles contributing the most to the @xmath7-axis conductivity are near the north and south pole . for the open fermi surface these quasiparticles are absent , and the @xmath7-axis conductivity is reduced by an additional factor of @xmath229 . it is also obvious , from comparing this with the results of the previous section , that the residual anisotropy can be close for the models c and d , and may not provide the unequivocal distinction between the two . the normal state scattering rate.,title="fig:",height=240 ] [ fig : dkap ] [ fig : drs ] it follows from experiment that the low-@xmath1 value of the anisotropy of the thermal conductivity is below 25% of the normal value . the analysis above shows that model a ( simple @xmath110-wave with constant interplane coupling ) does not give results that can explain the observed anisotropy . model b ( weakly modulated , along the @xmath7-axis , amplitude of the @xmath110-wave gap ) shows some tendency towards the temperature - dependent anisotropy , but , in our view , requires fine tuning if it were to explain the measurements of ref . . on the other hand , both model c ( modulated interlayer hopping , vertical line nodes ) and model d ( horizontal line nodes ) give very similar behavior of the thermal conductivity as a function of temperature , with a low residual value of the anisotropy @xmath8 , and therefore may be relevant to ceirin@xmath0 . distinguishing between the two based purely on the thermal conductivity data may not be easy , as is seen from the comparison in fig.[fig : comprs ] , which shows that , for sufficiently high anisotropy of @xmath230 , there is essentially no difference in the behavior of the ratio @xmath8 as a function of temperature between the two models . there are notable differences between the results presented for either model and the experimental observations of ref . . in experiment , the anisotropy ratio does not decrease below 0.5 for temperatures as low as 0.1@xmath14 , and there is no sign of saturation of the ratio below 0.15@xmath14 as in theoretical analysis . this is due to strong inelastic scattering that results in a peak in @xmath8 at about @xmath231 , but which is not included in our analysis here . correspondingly , the low-@xmath1 fit to experiment is done just below the peak , and the rapid decrease of the anisotropy may be related to the inelastic scattering . the low residual value of the thermal conductivity anisotropy implies , in model c , highly anisotropic tunneling , @xmath232 . published band structure data do not allow to reliably extract this ratio from fitting the quasi - two dimensional sheet of the fermi surface , but this is the task that perhaps should be attempted in near future . in the meantime , we propose several experiments that have the potential to provide additional information to resolve the question of vertical vs. horizontal line nodes . 1 . _ direct measurement of the specific heat and/or thermal conductivity in the vortex state under a rotated applied field . _ this is , in our view , the best test for the existence of the line nodes . if , for the field rotated in the @xmath34 plane , no difference in the specific heat is found across the @xmath233-@xmath1 phase diagram for the field direction along [ 110 ] vs. [ 100 ] direction , it is likely that there no vertical lines of nodes . if , on the other hand , such anisotropy is found and develops in agreement with theoretical predictions @xcite , this would be a strong evidence for @xmath110-wave , rather than hybrid , gap . + at the same time , it is possible , although , in our view , unlikely , that there is gap modulation _ both _ in the plane and along the @xmath7-axis . therefore measurements of the specific heat and thermal conductivity when the field is rotated in the @xmath234 or @xmath235 plane should be used to eliminate this possibility , and directly probe for horizontal line nodes . low superconducting transition temperature of ceirin@xmath0 and inelastic scattering require that these measurements be done at temperatures below @xmath236150mk . measurement of the @xmath7-axis thermal conductivity in cecoin@xmath0 . _ in the co compound there is a general agreement that there are vertical line nodes . measurements of the temperature evolution of the anisotropy @xmath8 in that compound , and comparison with ceirin@xmath0 , while short of proof , would provide a test for the connection between the anisotropy and the location of the nodes . note that there is some evidence for remaining unpaired electrons in the superconducting state of cecoin@xmath0 ( at least upon la - doping ) @xcite , which leads to a residual metallic linear term for both directions of the heat current . however , as it was suggested that the fraction of these quasiparticles can be determined experimentally @xcite , their contribution can be subtracted , and the remaining anisotropy used to test the differences and similarities between cecoin@xmath0 and ceirin@xmath0 . @xmath7-axis penetration depth measurements . _ in analogy to cuprates , if the interplane transport is suppressed along the nodal directions , the @xmath7-axis penetration depth should either be not linear in @xmath1 at @xmath20 ( for @xmath237 ) or have a small range of linear @xmath1 behavior , with the coefficient much smaller than that inferred from the density of states varying as @xmath238 . in cecoin@xmath0 the @xmath7-axis penetration depth is linear in @xmath1 , and observation of the non - linear behavior in ceirin@xmath0 would point towards differences between the two systems . unfortunately , we expect that , for reasons similar to those outlines for the thermal conductivity , it would be difficult to distinguish model c from model d on the basis of this measurement . negative result , however ( linear , in @xmath1 , @xmath239 ) would be difficult to reconcile with the thermal conductivity measurements . 4 . _ doping studies of residual @xmath7-axis thermal conductivity in ceirin@xmath0 . _ in fig.[fig : reskap ] we show the evolution of the residual thermal conductivity along the @xmath7-axis , in units of the expected universal value for that direction , as a function of the low - energy scattering rate , @xmath93 . notice that the increase in @xmath240 for the hybrid gap is slow , while for the modulated interplane hopping of model c it is faster . in reality we expect that addition of impurities will tend to make the interplane hopping more isotropic in addition to introducing pairbreaking , and therefore the increase in the residual value of @xmath240 will be even faster than that suggested by fig.[fig : reskap ] . we therefore predict that , if model c of vertical line nodes with modulated @xmath54 is realized , disordering the sample will produce a much more pronounced effect on the interplane thermal conductivity than that suggested by a hybrid gap . in conclusion , we investigated the constraints placed on the shape of the superconducting gap in ceirin@xmath0 by the results of shakeripour et al @xcite , assuming a one - band model with the fermi surface open along the @xmath7-axis of the crystal . we find that the temperature evolution of the anisotropy between the transport along the @xmath7-axis and in the plane is incompatible with a simple @xmath110-wave gap over a quasi - two dimensional fermi surface , such as often assumed in the studies of related compound , cecoin@xmath0 . the temperature variation of the anisotropy shows that the @xmath7-axis conductivity is affected more significantly than its intraplane counterpart by the opening of the superconducting gap . this implies that the ability of the unpaired quasiparticles to carry heat current along the @xmath7-axis is impaired relative to their contribution to the in - plane transport . this disparity can arise either from the gap anisotropy or the fermi surface properties . we explored several models where such modulation has different physical origin . we find that a weakly @xmath7-axis modulated @xmath110-wave gap does not easily provide satisfactory agreement with experiment , although may do so with fine tuning of the parameters . of the models with constant interplane hopping , that with * horizontal line nodes * most naturally account for the anisotropy , in partial agreement with the argument of shakeripour et al . remarkably , we also find that an alternative model , where the @xmath7-axis hopping depends on the direction of the in - plane quasiparticle momentum , yield results that agree with the experimental data even if purely @xmath110-wave gap , with * vertical line nodes * , is assumed . such an agreement requires substantial variation of the hopping between the nodal and the antinodal directions , in qualitative agreement with the analysis of the fermi surface obtained in the dhva measurements . given that there is some evidence for a distinct superconducting dome in ceirin@xmath0 it is clearly very important to determine the shape of the gap . we therefore suggested several possible experiments aimed directly at distinguishing the two situations and hope that further analysis will help resolve the uncertainty regarding the gap symmetry in this material . we also believe that the analysis above is generally relevant to a number of unconventional superconductors with a quasi - two - dimensional parts of the fermi surface . we are very indebted to h. shakeripour and l. taillefer for discussing their results with us before publication , and to m. a. tanatar for correspondence . we are also grateful to c. capan for critical reading of the manuscript . this research was supported in part by the board of regents of louisiana .
we analyze the behavior of the thermal conductivity in quasi - two dimensional superconductors with line nodes . motivated by measurements of the anisotropy between the interplane and intraplane thermal transport in ceirin@xmath0 we show that a simple model of the open fermi surface with vertical line nodes is insufficient to describe the data . we propose two possible extensions of the model taking into account a ) additional modulation of the gap along the axial direction of the open fermi surface ; and b ) dependence of the interplane tunneling on the direction of the in - plane momentum . we discuss the temperature dependence of the thermal conductivity anisotropy and its low @xmath1 limit in these two models and compare the results with a model with a horizontal line of nodes ( `` hybrid gap '' ) . we discuss possible relevance of each model for the symmetry of the order parameter in ceirin@xmath0 , and suggest further experiments aimed at clarifying the shape of the superconducting gap .
entanglement of two quantum bits ( qubits ) is a key feature to understand the microscopic world @xcite and also a basic resource of an important branch of quantum technology @xcite . entanglement is however a fragile resource . it is sensitive to a coupling of an entangled qubit to unavoidably present surrounding environment @xcite . the environment represents a class of quantum channels @xcite , which may reduce the entanglement or possibly even break it completely . in our analysis , we regard such channel as quantum if it is not entanglement breaking @xcite . we consider the most common incoherent environment containing multiple non - interacting and non - interfering qubits . in such a case , there is only a single mechanism reducing the entanglement . the qubit going through the environment can be lost and another fully incoherent qubit with noisy features can be found at the channel output . the incoherent qubits in the environment are considered as distinguishable from the qubits being transferred @xcite , but technically indistinguishable . it was proven that an incoherent environment containing only a single qubit is already sufficient to break entanglement in the channel @xcite and also substantially reduce the direct applicability in quantum technology @xcite . remarkably , if the entanglement remains after the channel , a conditional entanglement distillation can be used to approach the maximal entanglement and recover the quantumness of the channel @xcite . however , if the entanglement is completely lost , no entanglement distillation can help to fully recover the ideal quantum channel . the channel becomes entanglement breaking @xcite and the environment degrades the channel to classical one . the preservation of entanglement is therefore a fundamental limit for all the quantum channels in entanglement - based technology . to test this limit for the channel , we employ a maximum entanglement between a reference qubit r and the qubit entering the channel . at the output of the channel , we get qubit a leaving and verify the entanglement between qubits r and a. we consider the environment to be a large reservoir of qubits . as a result , we can represent the state of each environmental qubit by a mixed state @xmath0 . the diagonal basis consists of a ground state @xmath1 and an excited state @xmath2 , which is populated with a probability @xmath3 . for an environment at thermal equilibrium , the basis states become energy eigenstates with @xmath4 , where @xmath5 is a difference in energy between the ground and excited states of the qubit , @xmath6 is the temperature of the environment and @xmath7 is the boltzmann constant . first , a straightforward step is to find the maximum @xmath3 , or the maximum temperature @xmath6 of the environment , which still guarantees non - vanishing entanglement after the channel . this deterministic approach yields an unconditional limit on a cooling of the environment to keep the channel quantum . in the previous works @xcite , only single incoherent non - interfering and non - interacting qubit in the environment was analysed and the limit was experimentally tested . however , a realistic environment is typically complex , consisting of many non - interacting and non - interfering noisy qubits . therefore , this case is the main subject of our analysis here . . a projective measurement is carried out on output b , conditioning an entanglement preservation on output a. ] in this paper , we derive and experimentally verify a conditional cooling limit for quantum channel with incoherent environment with many independent noisy qubits . to derive the conditional limit , we extract an auxiliary qubit b from the incoherent environment , as is depicted in fig . 1 . it can be advantageously used to herald more entanglement between qubits r and a , thus beating the unconditional cooling limit . in this approach , the entanglement is fully broken only if the qubit entering the environment is lost and two completely incoherent environmental qubits appear at the outputs of the environment . the probability of the qubit being lost is @xmath8 . we derive a limit on a joint error represented by the product @xmath9 , depending on temperature @xmath6 for thermal environment . the limit is given by the probability @xmath10 of successful implementation of the ideal quantum channel . we experimentally verify this fundamental cooling limit using a quantum optics experiment with a simulated , controllable environment . however , it must be noted that our analysis is not limited solely to thermal environments or single photons ; it is applicable to any environment consisting of independent non - interacting and non - interfering qubits . the conditional cooling limit for qubit quantum channel is universal , widely applicable for different physical platforms . to derive an unconditional cooling limit , we ignore the possibility of access by any auxiliary output in the environment . we focus only on the case when a single qubit is present in the output a. an environmental qubit is considered in the state @xmath11 , where @xmath1 is the ground state of the environment and @xmath12 is the probability of a thermal excitation of the environment . to derive the maximum of the probability @xmath3 ( for thermal environment , a maximum of temperature @xmath6 ) , we consider a maximally entangled state @xmath13 between the reference qubit and the channel qubit , written in the basis of the environmental state @xmath0 . this state was chosen as a probe for the channel , because any other state would yield a more or equally strict condition on @xmath3 . as a result , the condition for @xmath3 represents a necessary criterion for the channel itself to be quantum , preserving entanglement . after the channel , the state @xmath14 changes to the unconditional state @xmath15 where @xmath10 is the probability of successful transmission of the maximally entangled state . the entanglement is preserved for any @xmath3 only if @xmath16 . otherwise , the general implicit condition on a quantum channel is @xmath17 which follows from using ppt entanglement criteria @xcite . when conventional cooling techniques are used to reduce the temperature of the environment , the limit of small residual @xmath18 becomes available . the inequality can be approximated by @xmath19 which is the very simple condition determining how well the environment has to be cooled down to still allow a quantum channel . for @xmath20 , the approximation is also valid . if the condition is satisfied , the condition is as well . the inequality is reminiscent of a condition widely used to verify nonclassicality of single - photon sources @xcite . here , @xmath10 stands for the success probability of transmitting a pure singlet @xmath14 through the environment , and @xmath3 is the measure of error , represented by a random thermal excitation of an environmental qubit . suppose now that an auxiliary qubit b , extracted from the environment , can be detected to herald the state in the outputs r and a. no coherent operation between the qubits b and r a will work since b needs not be coherent with r and a. a measurement on b is therefore the only way how to improve the unconditional limit . differently to the previous case , a random mixture of three elementary effects is present now in the environment . with a success probability @xmath10 , the channel yields the qubit entangled with r to the output a , so the singlet is transferred unchanged . this leaves a noisy state @xmath0 for the auxiliary output b. with a flip probability @xmath21 , the qubit entangled with r emerges from the auxiliary output b , while on a we find the noisy state @xmath0 . with a probability of loss @xmath8 , two completely incoherent qubits @xmath22 appear at the outputs a and b. this last process is extremely destructive , the entanglement with reference r is lost in the environment . all three effects are considered to be technically indistinguishable from outside of the environment . due to @xmath23 , we can use only @xmath10 and @xmath8 to fully characterize the resulting state @xmath24 describing a broad class of physical situations at many experimental platforms . to detect the state of qubit b , we assume general projective measurement @xmath25 . for any @xmath26 , the optimal strategy is to implement the projector @xmath27 on the more probable ground state . the resulting state is then @xmath28,\end{aligned}\ ] ] where @xmath29 . in this case , the conditional state preserves entanglement between r and a if the probability @xmath10 of success satisfies @xmath30 the joint error magnitude @xmath31 simply incorporates both undesirable sources of probabilistic error : the thermal excitation of the environment @xmath3 and the probability of entanglement loss @xmath8 . without either of these errors , the channel would always be quantum . the case of @xmath32 was predicted and already experimentally tested in refs . @xcite . however , it did not test an environment with more than a single qubit . in this case of @xmath33 , the projection on qubit b allows us to compensate @xmath8 by cooling the environment in order to fulfil condition . at a high temperature limit @xmath34 and for small @xmath35 , the condition @xmath36 substitutes the more strict condition @xmath37 for a guaranteed entanglement - preserving channel . more interestingly , for very cold environments with @xmath38 , the condition can be well approximated by the condition @xmath39 simply comparable to . the probability @xmath3 in is now substituted by @xmath31 , which becomes lower when the environment is cooled down . the condition represents the conditional cooling limit for quantum channel through incoherent many - qubit environment . experimentally , we have no way of directly measuring the channel parameters @xmath10 and @xmath8 , because we have no access to the noise process . we are , however , able to observe the preservation of entanglement . to positively recognize the limits of the quantum channel , we need the necessary conditions , to be sufficient as well . therefore , we use the maximally entangled singlet state as a probe , like we did in the theoretical analysis . in previous work @xcite , only single - photon noise was considered . this case is represented in our parametric space by the plane @xmath32 . our proposed simulator covers a more general case of non - zero @xmath8 . for our proof - of - principle measurement , we used the setup shown on fig . [ setup]a . the simulated parameters are then bound by @xmath40 . if one needs to simulate a general set of @xmath41 , one would simply use the environment shown on fig . [ setup]b . . as mentioned , the bounds can be tested in any basis . our experiment employed the polarization basis @xmath42 . a two - photon singlet polarization state @xmath43 was conditionally generated by a collinear type - ii parametric down - conversion in a bbo crystal . one photon was then propagated through a noisy channel , where a polarization state @xmath44 was incoherently coupled as noise on a 50:50 beam splitter . an attenuated laser diode was employed as the source of this independent noise . probing was done by splitting the signal on the second 50:50 beam splitter and detecting an auxiliary photon on detector det3 . eventually , both output ports r , a were subjected to a polarization projection . silicon avalanche photo - diodes ( excelitas ) were used for detection . for data acquisition , a time - tagging module ( qutools ) was used . we carried out a state tomography to reconstruct the state on the outputs r , a. every measurement was conditioned by all three detectors clicking . the coincidences between r and a filter out residual single photons present in each arm , that are inherent to every photon - pair source that includes optical loss . the detection on the auxiliary output b represents a successful extraction of a particle from the noisy environment . the polariser pol serves as a projective measurement @xmath45 . conditioned by a detection on output b. each dataset represents various temperatures of the noise for a certain noise intensity , which determines @xmath10 and @xmath8 . solid grey lines represent the unconditional limit and coloured dashed lines represent the conditional limit for the respective @xmath8 values . the space is limited by the condition @xmath46 , which visibly cuts both separability limits for higher @xmath8 . for a clear visualisation of numerical values , see fig . [ fig2d ] . ] , where each colour denotes a certain value of @xmath8 . square points represent the states , for which the entanglement was preserved only conditionally . ] the output state generated by the simulation is approximately of the form . in terms of experimental parameters , @xmath47 , where @xmath48 . @xmath49 is the count rate of the singlet state , @xmath50 the rate of single photons in output r and @xmath51 the rate of noise . @xmath52 is the coincidence window . in order to manipulate the two parameters @xmath3 and @xmath53 effectively , we performed a measurement for @xmath54 , separately . by probabilistically mixing the detections , we have control over @xmath3 without the need to carry out a full state tomography for each value . on the side of @xmath53 , the most convenient parameter to change is the coincidence window @xmath52 . in figures [ fig3d ] and [ fig2d ] , the measured data are shown . [ fig3d ] best illustrates the parametric space spanned by @xmath55 . the blue surface represents the conditional bound for separability , while the thick grey lines belong to the surface representing the unconditional bound . therefore , spherical data points below the blue surface represent quantum states , where the entanglement is lost . spherical points above the surface are the states , which remain entangled unconditionally . cube points are the states between the two conditions separable unconditionally , but entangled using the auxiliary projection . each line represents a certain kind of noisy channel parametrised by @xmath56 , experimentally set using the coincidence window @xmath52 . the points along the data lines correspond to various purity of the noise state @xmath0 , quantified by @xmath3 . the data illustrate that for sufficiently low temperature of the noise , here represented by the purity of @xmath0 , the noisy channel is not entanglement - breaking . in our demonstration , we used only two degrees of freedom , @xmath3 and @xmath53 , because we did not need to simulate any specific @xmath56 . to obtain the third degree of freedom , one needs to couple the incoherent noise to outputs a and b separately , with different intensities ( see fig . [ setup]b ) . assuming an ideal epr source , this allows covering the whole parametric space . for a realistic entanglement source , however , the regions of @xmath57 must be discussed . the first limit is the lower bound on the overall coupled noise intensity , which needs to be much stronger than single photons generated by the epr source . this approximation is necessary to experimentally obtain the state . as a result , one can reach the region of @xmath58 either by reducing inherent losses in the epr source , or by using nonclassical single - photon noise @xcite . the opposite limit of a strong noise gives the bounds @xmath59 and @xmath60 , where @xmath61 is the singlet generation rate relative to single - photon background in output r. in this case , if either @xmath62 , one needs only to sufficiently attenuate the signal before the environment to decrease the ratio @xmath63 . we presented necessary conditions to preserve entanglement propagated through an incoherent environment . we showed that entanglement can be preserved using a local auxiliary projection even for a many - particle environment . for sufficient cooling of a thermal environment , these conditions are reduced to simple error - success ratios , . we also presented a photonic experiment as a convenient way to simulate this scenario for an arbitrary set of parameters . finally , we experimentally verified that entanglement can be conditionally preserved using the proposed probing technique . the conditional cooling limit is also an interesting and relevant problem for other quantum channels coupled to a thermal environment . for a single - mode continuous - variable channels with thermal noise , a conditional correction restoring the entanglement - preserving nature of the channels was already proposed and experimentally tested in an all - optical simulator @xcite . recently , the conditional approach to ground state preparation was experimentally used to cool down a mechanical cantilever in a pulsed regime @xcite . the conditional preparation of entanglement between two mechanical systems was proposed in @xcite and is now being investigated experimentally . in future work , it would be therefore stimulating to extend the conditional cooling limit to quantum optomechanical systems . this work was supported by the czech science foundation ( 13 - 20319s ) . m. mikov acknowledges the support of palack university ( iga - prf-2015 - 005 ) . r. f. provided the theoretical concept and analysis . m. mikov ' a , m. j. , i. s. , and m. miuda performed the experiment . m. j. and m. d. supervised and coordinated the experiment . i. s. and r. f. wrote the main text . all authors were involved in editing and revising the manuscript .
we propose and experimentally verify a cooling limit for a quantum channel going through an incoherent environment . the environment consists of a large number of independent non - interacting and non - interfering elementary quantum systems qubits . the qubits travelling through the channel can only be randomly replaced by environmental qubits . we investigate a conditional cooling limit that exploits an additional probing output . the limit specifies when the single - qubit channel is quantum , i.e. it preserves entanglement . it is a fundamental condition for entanglement - based quantum technology .
simultaneous imaging in different filter - bands is of interest in a variety of astrophysical areas . the primary aim is to measure the spectral energy distribution or its evolution in variable objects in order to uncover the underlying emission mechanism . examples are , among others , ( 1 ) monitoring of all kinds of variable stars ( flare stars , cataclysmic variables , x - ray binaries ) to determine the outburst mechanisms and differentiate between physical state changes and changes induced by geometrical variations , like eclipses ; ( 2 ) monitoring of agn to understand the physical origin of the observed variability ; ( 3 ) determining the inclination of x - ray heated binaries @xcite ( 4 ) mapping of galaxies to study the stellar population ; ( 5 ) multi - color light curves of supernovae @xcite ; ( 6 ) differentiate achromatic microlensing events @xcite from other variables with similar light curves ; ( 7 ) identifying objects with peculiar spectral energy distributions , e.g. photometric redshift surveys for high-@xmath2 active galactic nuclei , or identifying brown dwarfs ; ( 8) follow - up observations of transiting extrasolar planets @xcite ; or ( 9 ) mapping of reflectance of solar system bodies as a function of their rotation to map their surface chemical composition @xcite . a new need for multi - band imaging arose with the observation of a large number of gamma - ray burst ( grb ) afterglows with the _ swift _ satellite @xcite . with its much more sensitive instruments it detects grbs over a very wide redshift range . since intermediate to high - resolution spectroscopy to measure the physical conditions of the burst environment ( e.g. * ? ? ? * ) is constrained to the first few hours after a grb explosion , a rapid determination of the redshift became important . this is best done with multi - band photometry ( until integral field units have grown to several arcmin field - of - views ) and deriving a photometric redshift based on the ly@xmath3 break @xcite . previous and current instruments with simultaneous imaging capability in different filter bands include andicam ( a novel double - imaging camera ; two channels , one for visual , the other for near - infrared @xcite , presently operated at a 1.3 m telescope ) , busca ( bonn university simultaneous camera ; four visual channels @xcite , presently operated at the 2.2 m telescope at calar alto ) , hipo ( high - speed imaging optical photometer for occultations ; two visual channels @xcite , to be operated on sofia , the stratospheric observatory for infrared astronomy ) , mitsume ( multicolor imaging telescopes for survey and monstrous explosions ; three channels with fixed bands g , r@xmath4 and i@xmath4 @xcite , operated at a 50 cm telescope ) , trispec ( triple range imager and spectrograph ; three channels with one ccd and two near - ir detectors and wheels for filters , grisms and wollaston prisms @xcite ) , sqiid ( simultaneous quad infrared imaging device ; @xmath1 and narrow - band @xmath5 filters in front of individual 512x512 quadrants of an aladdin insb array , designed for the f/15 cassegrain foci of the kpno 2.1-m and 4-m telescopes @xcite ) , ultracam ( ultrafast , triple - beam ccd camera for high - speed astrophysics @xcite ; portable instrument which has been used , among others , at the very large telescopes at eso , or the william herschel telescope , canary islands ) . here we describe the design ( @xmath62 ) and performance ( @xmath66 ) of a 7-channel imager , called grond ( * g*amma-*r*ay burst * o*ptical and * n*ear - infrared * d*etector ) , which was specifically designed for grb afterglow observations . we also mention some basics of the operation scheme ( @xmath63 ) , related software ( @xmath64 ) , and the changes to the telescope infrastructure which were implemented to use grond for rapid follow - up observations ( @xmath65 ) . the primary goal of identifying grb afterglows and measuring their redshift led to the concept of a camera which allows observations in several filters throughout the optical and near - infrared region at the same time . the simultaneity is dictated by the fact that a typical grb afterglow fades by about 23 mag between 510 min after the grb , and by another 4 mag in the following 50 min ( e.g. kann 2008 ) . rapidly determining the photometric redshift of grbs at @xmath7 requires multi - band photometry in at least 34 bands . in order to have a stable lever arm to determine ( i ) the intrinsic power law slope of the continuum emission and ( ii ) the galactic foreground as well as grb host intrinsic extinction , near - infrared bands up to @xmath8 are essential . extending the wavelength coverage beyond @xmath8 by including , e.g. an @xmath5 channel , was seriously considered during the design phase , but both the prospects of detecting a grb in the @xmath5 band as well as the substantial , additional technical constraints due to the required lower temperatures led us to drop this possibility . going as blue as possible is warranted by detecting the ly - break down to redshifts of @xmath93 ; however , also including a @xmath10 or @xmath11 band turned out to be difficult due to space and logistics constraints . leaving out one or two bands within our wavelength band would increase the error in the photo-@xmath2 determination by a factor of three at least ( for the redshift range covered by that band ) . thus , we decided to use four bands in the visual , plus the standard @xmath1 bands in the near - infrared ( nir ) . the field - of view ( fov ) of the camera should be large enough to cover the typical error boxes of gamma - ray bursts , but on the other hand have a pixel scale less than the mean seeing to allow accurate photometry . given the typical brightness of grb afterglows , a reasonably large telescope with permanent access is another important requirement . finally , as grond would act as pathfinder to pre - select `` interesting '' grbs for more detailed follow - up observations , a nearby 810 m telescope was considered advantageous in the search for the final host telescope . ] the final choice for the telescope was the mpi - owned 2.2 m telescope , operated by eso on la silla ( chile ) ; see @xmath6[tel ] . this telescope is a f/8.005 ritchey - chretien telescope @xcite with a focal length of 17600 mm , on an equatorial fork mount ; the intrinsic image quality is 04 . ] the prime goal for grond is to investigate the high - redshift universe , thus emphasis was put on the nir channels , and not on the visual ones . consequently , it was decided to develop a focal reducer in the nir with a 10@xmath12 10 field - of view in order to cover grb error boxes with a typical error radius of a few arcmin ( anticipated to occur about twice per week ) . in contrast , in the visual channels , no lens system was planned and the plate scale provided by the telescope optics was used . the separation of the different photometric bands has been achieved using dichroics , whereby the short wavelength part of the light is always reflected off the dichroic , while the long - wavelength part passes through . thus , the first dichroics are used to split off the visual bands which is done in the converging beam before the telescope focus . the relatively short back focal distance of 550 mm implied very tight constraints on separating the beams and still having enough space for the detectors . this , unfortunately , also prohibited using larger - format ccd detectors , since the consequently larger beams could not be separated further . after passing a diaphragm at the nominal telescope focal plane , the beam is folded by 180 for space constraints , and fed into a collimator ( fig . [ optlayout ] , [ 3dopt ] ) . the splitting of the three near - infrared channels occurs in the parallel beam between collimator and camera objective of the nir focal reducer . one might expect different ghosting effects in the visual arm ( dichroics in converging beam ) and the near - ir arm ( parallel beam ) . while it is true that reflections from the back surface of the dichroic in the parallel beam produce ghosts which are not offset relative to the main image , the effect of these ghosts is negligible . both the dichroic as well as the anti - reflection coatings have efficiencies of about 99.5% in their transmitting regions . since back surface ghosts involve two coating passages , they arrive with @xmath130.0025% ( corresponding to 11 mag fainter ) of the original intensity , undiscernable in data analysis . another typical concern of dichroics is their polarisation dependent location of the cut wavelengths , in particular since gamma - ray burst afterglow emission shows variable optical / nir polarisation . we have therefore embarked on a dedicated study of this effect prior to the coating of the optics . while standard cut wavelengths differ by 7% ( 3 - 4% ) for dichroic cubes ( plan - parallel plates ) between the perpendicular polarisation directions , the final grond coatings only have a difference of 1% ( e.g. 15 nm for the @xmath14 dichroic coating ) . for a 100% polarised white light beam this leads to a 0.1% change in the transmitted flux , which in turn implies a 10@xmath15% change in the color terms ( see @xmath66.2 ) , and thus is tolerable . the entire system has been designed for a working temperature of 80 k in vacuum , because deviations from the nominal temperature by more than 20 degrees would lead to refractive index changes large enough to cause noticable image quality degradation , and we considered it too risky to predict an equilibrium temperature gradient between the visual and nir part . since the mechanical assembly was done at room temperature , all optics ( and mechanical ) system parameters had to be calculated for a temperature of 80 k. for example , the focal length of the focal reducer differs by 3 mm between 293 k and 80 k , compared to a depth of focus of 25 @xmath16 m . we imposed a high accuracy requirement on the mechanical support structure , so that an alignment of the optics system is only required along the optical axis . for the visual channels the incoming 1:8 telescope beam is split using 4 dichroics ( top panel of fig . [ optlayout ] ) . the field of view is determined by the telescope focal length ( 17.6 m ) and the size of the ccds ( @xmath17 m ) . for each ccd the field of view is @xmath18 arcmin@xmath19 , and the plate scale is 0158/pixel . all four ccds are backside illuminated e2v devices without anti - blooming structures and are operated in inverted ( aimo ) mode . the chips come on nickel plated ceramic base plates . since each of the detectors serves in one and only one filter band , different sensitisations of the detectors have been applied : @xmath20 : astro - broadband on normal silicon , @xmath21 : mid - band on normal silicon , @xmath22 and @xmath23 : basic - nir on deep - depletion silicon . the use of dichroics implies that adjacent bands have identical 50% transmission wavelengths , making the sloan filter system @xcite the obvious choice . thus , the dichroics were designed such that the combination of their cut - off wavelengths defines bands identical to the sloan system , with the exception of the @xmath22 band . since in the sloan system @xmath24 overlap at their @xmath2570% transmission values , we decided to compromise the @xmath22 band in favour of standard - width @xmath21 and @xmath23 bands . channel only involves reflections from the dichroics , and therefore is symmetric ( and thus shows the image quality provided by the three mirrors of the telescope ) . each panel is 27@xmath16m@xmath1227@xmath16 m , corresponding to 2x2 pixels of the detector . [ visspot ] ] the telescope beam is first split by reflecting off the @xmath20 and @xmath21 bands , then the @xmath22 and @xmath23 bands . the long wavelength part of the beam ( the nir part ) is then passed through the primary cold stop in the focal plane . two further dichroics are placed into each of the two visual beams to separate @xmath20 and @xmath21 , and @xmath22 and @xmath23 , respectively . the dichroic plates are wedge shaped by @xmath251319 arcmin so that they compensate for the astigmatism of a plane - parallel plate in the converging 1:8 beam of the telescope @xcite . the curvature of the focal plane of the telescope has a radius of 2205 mm , which corresponds to a focus difference of 43@xmath16 m between the center and edge of the ccd . this is smaller than the depth of focus of @xmath2560@xmath16 m at the shortest wavelength used , but of course causes astigmatism ( see fig . [ visspot ] ) . originally , no other optics were planned except for the dichroics . after first light it became apparent that the @xmath20 and @xmath22 bands suffered from single reflections of @xmath21 and @xmath23 light , respectively , off the backside of the dichroics , at a strength of about 0.5% of the incident intensity . thus , each of the ccd channels has been equipped with a short - pass filter to block these single reflections . the backsides of those 3 mm glass filters have anti - reflection coatings at @xmath2699.5% efficiency , except for the @xmath20-band ( see below ) . another post - commissioning change has been implemented for the @xmath20 band : originally , no short - wavelength cutoff was foreseen , since we expected the transmission of the telescope mirrors and the entrance window as well as the efficiency of the ccd detector to fall off rapidly enough below @xmath25400 nm . in fact , the ill - defined cut - off led to a photometric uncertainty of about 5% which was deemed unacceptable ( see @xmath6[limmag ] ) . we therefore implemented a short - wavelength edge - filter with a cut - off wavelength of 380 nm . the nir part is designed as a focal reducer system allowing for an optimized adjustment of the focal length , or equivalently the field of view . in this way the required 10 fov is imaged onto a @xmath27 rockwell hawaii-1 array ( pixel size 18.5@xmath16 m , plate scale of 060/pixel ) . the focal reducer system ( 5 collimator lenses with a resulting f@xmath28=360 mm and 6 camera lenses with f@xmath29=129.6 mm for each nir channel ) has a pupil reduction factor of 48.89 and a pupil diameter of 45 mm . for the f/8 telescope the resulting effective focal length is 6336 mm . a diaphragm is placed in the telescope and collimator focus , including blocking of the m2 mirror and m2 spider structure . the two beam splitters for reflecting off the @xmath30 and @xmath31 bands are located in the parallel beam of the focal reducer , between collimator and the camera objectives ( fig . [ optlayout ] , [ 3dopt ] ) . in addition , each channel has a cold pupil stop near the dichroic , as well as a filter in front of the first camera lens defining the standard @xmath1 bands ( which also prohibit long - wavelengths single reflections to reach the detector ) . the @xmath1 filters differ from canonical nir filters in their much reduced blocking range which is sufficient for grond because of the action of the dichroics . thus , the mean transmission in the @xmath1 bands ( including dichroics ) is 98% . because of the parallel beam , the dichroics could be manufactured as true plane parallel plates and mounted in a compact 60 box which allows for a space - saving splitting of the beam into the @xmath32 and @xmath8 channels . the @xmath8 channel includes an additional flip mirror for dithering purposes . but for the three infrared channels . the focal reducer was designed to optimize the image quality ( point source shape ) , but still contains image distortions , i.e. plate scale changes depending on the position in the focal plane . these geometrical distortions are corrected during the data reduction process ( see @xmath6[pipeii ] ) . each panel is 37@xmath16m@xmath1237@xmath16 m , corresponding to 2x2 pixels of the detector . [ irspot ] ] the focal reducer system designed for grond differs substantially from that of other optical focal reducer systems ( e.g. the calar alto faint object spectrograph cafos , the multi object spectrogragh for calar alto mosca , or the potsdam multi - aperture spectrophotometer pmas @xcite at calar alto ) , since no optical glasses could be used . basically only four media are available for transmission in the 9002400 nm range : caf@xmath33 , baf@xmath33 , oh - free silica and undoped yag . because of its transmission characteristics , yag is particularly suited for nir optics ( but also difficult to grow in large dimensions ) . using these four optical media , and after measuring their refractive indices at the foreseen operational temperature of 80 k , the focal reducer system could be designed with a high performance ( fig . [ irspot ] ) using software developed by us ( laux 1999 ) . the 3d visible and nir model as well as the optics components cad files were created using the zemax software package . the grond cryostat ( fig . [ cryostat ] ) consists of the following components : a ) grond - cylinder , made of titanium , and electron beam welded ; b ) the front and back plates made of aluminum ; c ) 4 spacer rods ; d ) 8 belleville - spring systems ; and e ) o - ring seals ( viton ) . ] the components inside the grond cryostat have different thermal requirements . in addition , in the case of single component failures , no sensitive components should be exposed to dangerous thermal conditions . the complete internal structure is mounted on a titanium alloy ( tialv ) plate that provides thermal insulation . this plate is mounted on the front plate of the cryostat , which accordingly becomes the nominal reference point of the system . as the different components ( front plate , titanium plate , optical bench ) have different temperatures , the mounting must allow for thermal contractions . this is achieved by using long holes and centering pins to mount the titanium plate . the internal components ( optical benches ) are connected to the back plate of the cryostat using a metallic membrane that allows contractions only along the optical axis of the instrument . the cooling system is based on a two - stage sumitomo closed - cycle cooler ( ccc ) with a srdk-408-s cold head which provides a cooling power of 35 w down to 45k at the warm stage , and 6.3 w down to 10k at the cold stage . the cold head is mounted using a bellow ( to reduce vibrations ) on the back plate of the vessel . the first stage of the ccc ( higher temperature and higher cooling power ) is used to cool the optical benches ( visual and near - ir ) and all optics components to 80k ( temperature - controlled ) . the second stage ( lower temperature and lower cooling power ) is used to cool the near - ir detectors to 65k ( optimum temperature to maximise the signal - to - noise ratio ) . the ccd detectors have an intentionally bad thermal connection to the visual optical bench ( 80k ) . they are stabilised at 165k using resistor heaters . thermal connections between the two baseplates and the components are realised using copper flanges , copper ledges and flexible copper lines . galvanic isolation of the cooling system is achieved by isolation foils being mounted between the thermal copper flanges and the ccc stages . to reduce the thermal input , the optical benches are thermally insulated from the cryostat . in addition , thermal shields are installed to reduce the radiative heat transfer . the main radiation shield is between the optical benches and the vessel . it is thermally insulated from both . the radiation shield is constructed of three round plates ( two at the front plate , one at the back plate of the vessel ) , four half cylinders around the spacing bars and six dismountable cylinder segments all made of 1.5 mm thick highly reflective aluminum plates . the cylinder segments are additionally covered with multi - layer insulation foil . ] the implementation of the optical requirements with the geometrical and mechanical constraints such as tight space , minimum weight and different temperature regions within the main housing resulted in substantial complexity which in turn required most sub - units and their interrelations to be treated simultaneously . the development of the optical / mechanical configuration included considerations such as : ( i ) space requirements for the 4 ccd cameras and 3 nir - detector units , ( ii ) use of structural components , such as optical benches , mounting interfaces / connections , and cold connections , ( iii ) assembly and handling procedures , ( iv ) cabling , ( v ) low - temperature induced contractions interacting in 3 dimensions . the final structural chassis ( fig . [ chassis ] ) consists mainly of two separate optical benches . the top bench ( fig . [ visbench ] ) contains the optical and mechanical groups for the visual light detection , in particular : ( 1 ) the entrance light baffle ( not shown in fig . [ visbench ] ) ; ( 2 ) principal light splitter unit with two dichroics ; ( 3 ) small dichroics splitting the @xmath20 , @xmath21 and @xmath22 , @xmath23 channels , respectively ; ( 4 ) two vis - shutter units for ccd - channels @xmath34 and @xmath35 , respectively ; ( 5 ) all four ccd - boxes for @xmath0 ; ( 6 ) nir channel shutter to close the nir sector completely ; ( 7 ) mirror carrier unit for folding mirror 1 ( also not shown in fig . [ visbench ] ) . the lower bench ( nir - area ; fig . [ irbench ] ) contains the following components : ( 1 ) 2nd folding mirror ( adjustable in 3 degrees of freedom ) ; ( 2 ) collimator lens unit ; ( 3 ) two nir beam splitters to separate @xmath30 , @xmath31 , and @xmath8 ; ( 4 ) one camera lens system ( split into two mechanical lens holders , fig . [ optlayout ] ) ; for each of the 3 nir channels ; ( 5 ) dithering mirror unit for the @xmath8-channel ; ( 6 ) nir - detector units ( motorised for focussing ) ; ( 7 ) zeolith boxes near the ccc cold stage as water ( humidity ) sinks ( not shown in fig . [ irbench ] ) . the nir light path is only open at the focal plane , but otherwise completely closed by a dedicated light - tight housing cooled to 80k . cabling is done through light traps . within the focal reducer , 23 lenses are maintained at precise position during the cooling to 80k . a special spring - based lens mounting design was developed ( fig . [ lens ] ) which keeps the lenses centered and in the correct axial position . during the cooling the radial flexibility of the lenses is kept small enough to ensure the required centricity . ] the tight space constraints also made it necessary to develop a new , much smaller detector board ( shrinking from originally 220 mm diameter to a square of 50x70 mm ) . consequently , the mechanics also had to be re - designed from scratch and the opportunity was taken to implement two new features : ( i ) motorised positioning of the detector along the optical axis , and ( ii ) a gimbal mounting to allow manual alignment of the detector plane perpendicular to the optical axis in both degrees of freedom ( fig . [ irunit ] ) . a cold finger on the backside of each nir detector is directly connected with the second stage of the ccc , thus allowing the nir detectors to cool below the temperature of the baseplate and optics . since the ccc temperature is load dependent , a heater is installed on the coldfinger to control the temperature and keep variations below @xmath360.1 k. ] the ccd detector units have no motorised focusing stages , because the depth of focus is larger . still , they can be aligned in 4 degrees of freedom ( fig . [ ccdunit ] ) . since the minimum allowed operational temperature of the ccds ( @xmath25155k ) is much higher than the inner cryostat temperature , the ccds are equipped with a heat finger and thermally controlled via the pulpo electronics ( see section [ pulpo ] ) to 165k within @xmath361 k. ] there are several motorized mechanisms in the instrument which operate at 80 k. all motors are from phytron , but use different gears . * _ nir channel shutter _ : this door closes the near - ir arm of the instrument , so one can take darks in @xmath30 , @xmath31 and @xmath8 . it is implemented using a stepper motor . * _ visual shutters _ : due to the constraints of the optical design of the instrument , the ccd shutters must be cryogenic ( as the shutter must not block the infrared channels ) . due to the tight space constraints , one shutter per ccd was mechanically impossible ; thus two shutters have been implemented , one for the @xmath34 and one for the @xmath35 channels , respectively . each shutter is a thin metal disc with two quadrant shaped openings at opposite sides that are slightly larger than the beam at this position . thus , each disc has two shutter closed and two shutter open positions , i.e. 4 possible positions in total which are 90@xmath37 apart . each shutter is equipped with a stepper motor and a position encoder that provides the ( absolute ) position of the disc . the combination of stepper motors and encoders ensures that movements are fast ( with stepper motors one can blindly move the shutter by 90 degrees ) and that the shutter does not drift over time , resulting in vignetting ( i.e. the 4 nominal positions are tied to encoder positions ) . this design is still slightly slower than typical , non - cryogenic shutters , but it is fast enough ( below 1 sec ) for typical exposure times of 1 min or more . the two shutters can be operated independently , thus , different integration times in the two optical arms are possible . shutter movement is triggered by the pulpo shutter control port ( see section [ pulpo ] ) . * _ flip mirror zero position _ : the flip mirror ( @xmath8-band channel dithering ) can be either in the calibration / alignment position ( i.e. no dithering ) or in the normal position ( where dithering is possible by rotating the flip mirror ) . this function is implemented using a stepper motor that moves the mirror into either position . * _ flip mirror dithering _ : this motor is responsible for the internal dithering in the @xmath8-band . dithering in @xmath8 as opposed to the other channels is required due to the high sky brightness . in order to allow staring at the same sky position in the optical bands ( because of the long ccd read - out times ) , this internal dither mirror has been implemented . the dither function is implemented using a cld stepper motor linear drive from phytron and an encoder . two fundamental operational modes are implemented . either the motor is moved to an absolute position or the motor moves in either direction 45 , 60 or 90 degrees ( i.e. 8 , 6 or 4 dither positions ) . * _ ir focus _ : each of the three nir detectors can be focused independently using 3 stepper motors with gears . each motor allows relative movements to the required accuracy of @xmath3625 @xmath16 m ) , over a range of @xmath362 mm . the gear is strong enough to prevent any shift of the nir detector with positioner in the best - focus position during telescope movements . the electronics are divided into several parts : ( i ) the main rack which contains the grond - specific control , ( ii ) the read - out electronics for the ccd and nir detectors as provided by eso . _ interface unit and main control _ this unit forms the control interface between the grond instrument workstation and the motor control electronics . the main part of this unit is a xilinx fpga , which makes the electronics more flexible . each io of this xilinx fpga is driven by a bidirectional bus driver and additionally esd protected by a low capacitance diode array . the current design uses 98 pins for input and 66 pins for output purposes , with 32 spare pins . * communication interface : * the communication interface is realised by an onboard rs232 interface with two jumpers for speed selection . the nominal operation speed is 115200 baud . a rs232 server converts this signal to 100baset for the instrument workstation . to maximise reliability , a special command structure is used . after the detection of the sync byte , the interface unit starts a timeout counter . if a command is not completed within 8 ms , this command will be rejected and the whole communication unit reset , so that this unit can receive new commands immediately . this protects the system from hanging due to incomplete commands . * motor control : * the main control contains all electronics necessary for the digital control of the motor driver boards and position readout . all signals which are important for operation of the motor driver electronics ( clock , boost , direction , reset ) , are delivered by the main control . for all position - switch controlled stepper motors ( ir - calibration , cold shutter , zero position ) , the digital circuit prevents a motor overrun ( i.e. an end switch ) . this means that as long as the stepper motor is in between the two end - switches the direction will be determined by the user . if one of the end - switches is reached , the motor stops immediately and the direction is automatically reversed . likewise , these electronics are responsible for the readout of the resolver values and synchronisation with the stepper motors . three resolvers are controlled by the grond electronics ; one for the flip mirror and two for the ccd shutters . an additional offset register can be used to calibrate the start position of the flip mirror . the movement is bidirectional and commandable . the accuracy of the flip mirror position is about 5.4 arcmin ( 4000 steps per full rotation ) . contrary to the flip mirror , the ccd shutter electronics have only one mode ( 4 @xmath12 90 degrees ) . the shutter is commandable over the pulpo ( see section [ pulpo ] ) interface , which gives a single bit state to the shutter electronics that determines whether it is open or closed . a commandable offset register for each ccd shutter helps the user to adjust the initial open / close position . the accuracy of the shutter is also 5.4 arcmin , corresponding to 80 @xmath16 m in the center of the opening . _ motor driver boards for stepper motors _ the motor driver board has two independent driver units . these boards are used for driving the nir focusing stage , the cold shutter and the zero - position motors . each driver unit has four current modes ( boost , run , standby , off ) , which are selected via two status bits . the current levels for the modes can be set using variable onboard resistors . additionally , each driver unit has its own temperature sensor , which prevents overheating . _ motor driver boards for linear motors _ this board was specially developed for high current with linear motors . linear motors are used for moving the m3 mirror and the main shutter door in front of the entrance window . the board works with two limit switches , between which the motor can move . the forward backward movement of the mirror is possible by software - controlled command ( from the grond instrument workstation ) or manually over a special control box at the telescope s m1 mirror cell . _ resolver boards _ an ad2s82a represents the main part of the resolver board , which is a variable resolver - to - digital converter . the resolution of this converter is variable and can be set up to 16 bit , which means that a maximum resolution of approximately 20 arcsec is possible . the sine - wave for the ltn resolver is produced by the maximum high frequency waveform generator max038 . this produces an optimal signal for the resolver . a high speed voltage follower is connected at the outlet side to provide a high current driver for long cables . likewise the offset of the sine - wave can be adjusted with a voltage buffer . all resolver boards used in the grond electronic rack have the following configuration : resolution : 16 bit ; peak - to - peak sine wave amplitude : 2 v ; frequency : 10khz ; offset : 0v . finally , the primary power of all motor power supplies is switched on by a zero crossing switch . this prevents a high inrush current on the main power line during switch on . _ temperature control _ the electronic rack has two temperature control systems . the temperatures of the nir detector system is monitored by lakeshore temperature controllers . each of the three nir detectors has its own lakeshore controller ls331 which in turn has two control loops : loop1 with 50w maximum heating power is used to warm up and regulate parts of the baseplate . loop2 with 1w maximum heating power controls the detector temperature with an accuracy of @xmath360.1deg . the nominal operating temperature of the grond nir detectors is 65k . for the optical ccds , temperature sensors and controlled heaters are directly mounted to the ccd detector backside and controlled via pulpo ( see section [ pulpo ] ) . since each pulpo can control only one shutter , two pulpos are incorporated in the electronic rack . the nominal heating power of the pulpo is 7w in each circuit . for safety reasons , the power at the ccds was reduced to 1w . a third heater , also controlled by each pulpo , is mounted on each of the two clock - bias board thermal shields , with 7w heating power . * irace : * the readout system for the infrared detectors , the irace ( infrared detector high speed array control and processing electronics ; meyer 1998 ) system , was supplied by eso . as read - out mode ( ` imode ` ) , the double - correlated mode is used , as it is the only valid mode for science observations . one change was made to the basic irace system , namely to read out all three detectors simultaneously and store the data into one fits file . * fiera : * the fiera ( fast imager electronic readout assembly ; beletic 1998 ) system was also provided by eso . implemented changes include ( i ) independent read - out for two pairs of ccds ; ( ii ) a 3.8 m instead of the nominal 1.8 m cable to the front - end electronics ( which does not increase readout noise ) . for grond , the two - port readout of each detector is used and one fits file with four separate extensions is written after each exposure ( the two read - out ports are combined into one image before writing to file ) . furthermore , two read - out speeds are typically used : the fast mode ( 625 kpixel / s ; read - out noise of @xmath923 e@xmath38 ) in the 4 min and 6 min observation blocks ( see below ) , and the slow mode ( 50 kpixel / s , @xmath95 e@xmath38 ) in all longer observation blocks . the fiera system includes pulpo @xcite , a separate electronics box used to measure and control temperatures ( via 3 heaters ) , measure the vacuum in the cryostat ( two inputs ) , and to operate one shutter . the grond operation scheme is designed to be as generic as feasible , but primarily suitable for prompt , automatic grb follow up observations . additionally , it should provide the basic functionality to perform observations of any other , non - transient object . however , operating a seven channel imager in a single cryostat implies various observational constraints which should be considered in the creation of the observation blocks ( obs ; chavan 2000 ) . each observation is built by combining exposures ( differing in number and length ) of the seven detectors based on an appropriate set of parameters . this approach provides the most generic background for science observations with the current instrument design . the most important parameters are the number of telescope dither positions ` ntd ` , the number of telescope pointings ` ntp ` , the number of exposures in the @xmath20 and @xmath21 band ` ngr ` , the number of exposures in the @xmath22 and @xmath23 band ` niz ` , the corresponding integration times ` uitgr ` and ` uitiz ` , the number of @xmath8 mirror dither positions ` nmd ` , the number of @xmath8 mirror pointings ` nmp ` , the number of @xmath1 exposures in a single @xmath8 mirror position ` nint ` , the number of stacked @xmath1 exposures ` ndit ` , and the @xmath1 integration time ` dit ` . in order to comply with the grond science objectives , different default ob types are defined for autonomous observations of grb afterglows . the obs are named after the total integration time in the @xmath8 band , namely 4 min , 8 min , 20 min and 40 min obs , all with 4 different telescope dither positions , as well as 6 min , 12 min , 30 min and 60 min obs with 6 telescope dither positions . all parameters for grond observations can be adapted for use in non - standard obs , within their constraints as described above . however , all bands and their readout have to be synchronised for telescope dithering , which means that the integration time of the optical bands must be adapted to the parameters of the infrared observations ( and/or vice versa ) . for non - standard obs , this remains the responsibility of the user . by default , one standard ob consists of four or six telescope dither positions : ` ntd = ntp=4 or 6 ` , each with six @xmath8 band mirror dither positions ( ` nmd = nmp=6 ` ) . a total of ` nint`=1 , 2 , 5 and 10 single @xmath30 , @xmath31 and @xmath8 integrations are possible at each @xmath8 mirror dither position . a generic break - down of the shortest 4 min ob and the underlying structure in a single telecope dither position can be seen in fig . [ fig : grondcam ] . schemes for the other default obs differ in the number of @xmath30 , @xmath31 and @xmath8 integrations ( ` nint ` ) and telescope ditherings ( ` ntd ` ) , and the ccd detector read - out mode ( ` omode ` ) . the grond instrument software is based on the standard eso / vlt software environment . the ccd detectors are handled as a normal array of 4 chips , each with 2-port read - out . the single fiera controller drives all 4 chips . the two shutters are operated in parallel . in order to integrate longer in the bluer ( @xmath34 ) channels , two special read - out modes were introduced : one that drives only the @xmath35 ccds ( i.e. one can read out only these two chips while leaving the image in the other two intact ) and a second read - out where the @xmath34 chips are not wiped / reset at the start of the exposure . when combined , these two readout modes allow a special operation mode where one takes one @xmath34 image for every two @xmath35 frames . thus , in the bluer channels the efficiency is significantly improved due to the additional exposure time . there is a resulting trade off between the dark time ( between reset / wipe and read - out ) and the exposure time in the @xmath34 channels . the latter can not be precisely known because of the not exactly predictable overhead of a fiera exposure . as the dark current is negligible in the ccds , this trade off is justified . for very short exposures of very bright targets a fast read - out mode is also provided , which reduces the read - out time from 46s to 4.4s . the near - ir chips are operated normally . the 3 chips with 4 amplifiers each are treated as 12 video channels and are stored in a single frame . this implies that all three chips must be read out together . as a consequence , the efficiency of the @xmath14 channels are slightly reduced : this is because the integration time for faint targets is limited by the sky brightness in the @xmath8-band to @xmath2510 seconds while the other two bands could , in principle , integrate longer . this results in a 2 second overhead . in addition , the @xmath14 chips must also wait for the dither mirror movements resulting in an additional 0.5sec overhead for each mirror movement . there is no additional read - out noise due to the larger number of @xmath14 reads since the total noise is dominated by poisson noise of the sky background . the two arms ( optical and near - ir ) are operated in semi - parallel . both detector systems are started independently consecutively ( within milliseconds ) and the requested number of frames are taken with as little dead time as possible . in general the two detector systems are _ not _ synchronised except after telescope slews . the instrument control software has no provisions to fill the time available between telescope movements . it is the user s responsibility to request the optimal number of exposures . in order to ease this task , a few optimal configurations are offered . relatively frequent telescope offsets are required to be able to determine the sky variations in the @xmath30 and @xmath31 band ( the @xmath8-band is handled by the built - in dither mirror ) . in order to increase efficiency and to simplify usage , the dither pattern is a predefined circle the user only defines the number of dither positions . grond uses so - called combined offsets , i.e. guiding continues without operator interaction after a telescope offset . the operator is only required to select the guide star at the beginning of an ob . observations , both in visitor and service mode , are done through obs . the user can use the normal p2pp tool ( p2pp manual 2007 ) like for any eso instrument ( despite grond not being an official eso instrument ) . in the case of grb observations , obs are generated in real time by a special process . the input parameters are provided by the grond pipeline host ( see next section ) , which is translated into a standard ob . to reduce overheads , the instrument workstation always provides the current telescope pointing to the pipeline machine . thus , the pipeline may conclude that the current pointing is good enough ( i.e. the aim coordinate and error circle fall within the field - of - view of the optical ccds ) and skip the telescope preset . in this case guiding will continue without operator interaction , even when executing several obs consecutively ( guiding does not stop at the end of an ob ) . at the start of each ob , the instrument is focused by moving m2 ( see @xmath6[foc ] . as grond is fine - tuned for grb follow - up where speed is crucial , no manual focusing is foreseen before a grb observation starts but the necessary templates are available for standard observing runs . to make data analysis easier , all images produced by grond contain important information in the fits headers . thus , the images can be fed into standard software such as iraf as the required header cards describing the bias / trim / overscan regions and amplifiers are included . the visual channels are also identified by the header ( @xmath0 ) ; the infrared channels are in one image ( see above ) . obviously , all other relevant parameters ( gain , read - out noise , etc . ) are also included for each chip in the header . crude astrometry ( wcs cards ) is also included in each image . in the optical the astrometry is correct , apart from a shift introduced by the pointing of the telescope . the astrometry is correct for all chips ( i.e. individual chip shifts , rotations , etc . are included ) . as there are no significant distortions in the optical channels , the astrometry is very accurate ( 02 ) once the pointing origin is fixed . in the near - ir channels , the astrometry does _ not _ include the focal plane distortions ( which is significant due to the focal reducer ) . furthermore , the near - ir detector system produces a single image so an extra processing step is required to slice the image into the three bands and to extract the necessary information from the header . the header cards also take into account the @xmath8-band dithering mirror position . [ cols="<,<",options="header " , ] a first set of calibration observations was performed during mid - july 2007 , though not in photometric conditions . thus , the constants in the following equations may change slightly ( see www.mpe.mpg.de/~jcg/grond for updates ) . first , we derive transformation equations between the grond ( g ) visual bands and the sloan ( s ) filter system : @xmath39 + @xmath40 + @xmath41 + @xmath42 next , the transformations between the grond nir bands and the 2mass ( 2 m ) system are as follows : @xmath43 + @xmath44 + @xmath45 for transformation from 2mass to other systems see carpenter ( 2003 ) . finally , the transformations between the visual grond bands and the wfi ( w ) system are : @xmath46 + @xmath47 + @xmath48 + @xmath49 only for the @xmath21 band we reach a residual 1% rms scatter , while for @xmath22 and @xmath23 the residual rms scatter is 2% . for a transformation to the standard landolt ( johnson @xmath50 and cousins @xmath51 ) system see the dedicated wfi web - page . grond is an imaging system capable of operating in seven colors simultaneously . it has been designed and built at mpe garching , and commissioned at the 2.2 m mpi / eso telescope on la silla , chile . first observations show that all properties are according to specifications / expectations . the first observations of gamma - ray bursts with grond have also been obtained ( greiner 2007a , b ; primak 2007 , krhler 2007 ) . fine tuning of the operations strategy as well as scheduling and analysis software in the upcoming weeks is expected to bring grond into a fully operational condition , thus allowing the commencement of normal science operations . we thank k. meisenheimer , r .- rohloff and r. wolf ( all mpia heidelberg ) for their support in getting the telescope interfaces right , and for fe computations of the m3 mirror rigidity . particular thanks to the whole la silla observatory staff for their enthusiasm and effort during the assembly of all the grond components to the telescope . we thank d.h . hartmann ( clemson univ . ) for stimulating discussions , k. garimella ( formerly also clemson univ ) and d.a . kann ( tautenburg observatory ) for help in implementing sed - fitting routines as preparatory steps for hyperz , as well as a. rossi ( tautenburg observatory ) for support in the derivation of the photometric conversion equations . we are greatful to the anonymous referee for the many detailed comments which improved to readability and consistency of the paper . we greatly acknowledge the special efforts of the following companies to fulfil our often unusual requirements : berliner glas kgaa , ecm moosinning , fee gmbh idar - oberstein , korth kristalle gmbh altenholz , laseroptik gmbh garbsen , laser zentrum hannover e.v . , przisionsoptik gera , steinbach - knitzer - lopez jena , tafelmaier dnnschichttechnik rosenheim . part of the funding for grond ( both hardware as well as personnel ) was generously granted from the leibniz - prize ( dfg grant ha 1850/28 - 1 ) to prof . g. hasinger ( mpe ) . meyer m. , finger g. , mehrgan h. , nicolini g. , stegmeier j. , 1998 , in proc . `` infrared astronomical instrumentation '' , ed . a.m. fowler , spie vol . 3354 , p. 134 ( see also www.eso.org/projects/iridt/irace )
we describe the construction of grond , a 7-channel imager , primarily designed for rapid observations of gamma - ray burst afterglows . it allows simultaneous imaging in the sloan @xmath0 and near - infrared @xmath1 bands . grond was commissioned at the mpi / eso 2.2 m telescope at la silla ( chile ) in april 2007 , and first results of its performance and calibration are presented .
after a brief , intense , and optically obscured protostellar phase , newly formed stars emerge from their natal environments as relatively bright , distended objects undergoing gravitational contraction towards the main sequence . while residual accretion from circumstellar disks is observed to take place in young pre main - sequence ( pms ) stars , the magnitude of the accretion flow appears to be much lower ( typically , @xmath1 m@xmath0 yr@xmath2 ) than the protostellar one ( typically , @xmath3 m@xmath0 yr@xmath2 ) . therefore , pms derive most of their energy from contraction rather than accretion , and evolve on a characteristic kelvin - helmoltz time scale ( @xmath4 ) . the major imprint left from the protostellar phase on pms stars is their specific initial conditions . at the end of the main accretion phase , the stellar core first appears as an optically visible star along the so - called birthline ( @xcite ) . empirically , the birthline provides the initial conditions for pms contraction , specifying the internal structure and radius of the newly formed star . depending on the mass of the star , young visible stars are classified as t tauri ( low - mass , typically @xmath51 - 2 m@xmath0 ) , or herbig ae / be stars ( intermediate - mass , @xmath510 m@xmath0 ) . for @xmath3 m@xmath0 yr@xmath2 , the birthline joins the zams at m@xmath6810 m@xmath0 ( @xcite ) . thus , for stars with masses above this value there should be no optical pms phase : the rapid gravitational contraction of massive protostellar cores has caused hydrogen ignition already during the accretion phase . therefore , massive stars are expected to be born _ on _ the zams . this brief excursus highlights the physical link between protostellar and pms evolution and justifies the notion that young massive stars have a distinctively different early evolution than the lower mass counterparts . to date , there is still a lack of numerical models capable of following these two phases self - consistently . while the powerful simulations described in these proceedings ( see the contributions by bonnell and klessen ) can follow the fragmentation , collapse and accretion flows of dense gas within cluster - like conditions , they can not resolve the innermost regions where the protostellar core grows in mass and eventually turns into a pms object . thus , in order to describe the main features of stellar evolution before the zams , we have still to rely on calculations that either ignore completely the protostellar phase as in the classical models of pms evolution ( e.g. , @xcite , @xcite ) or inherit the initial conditions from independent simplified treatments of the accretion flow . in the following , i will summarize the main properties of the latter class of models , referring to @xcite for a full discussion . then , i will discuss separately some important tests of the predictions of pms models covering the full mass spectrum . a key property of protostellar evolution is the existence of a _ mass - radius _ relation for accreting cores . this relation finds its physical origin in the thermostatic nature of deuterium burning once the protostar reaches the critical ignition temperature ( @xmath7 k ) . as a result of deuterium burning , first at the center and then in a subsurface shell , protostars never attain large radii : the radius remains typically 35 times larger than the corresponding value on the zams , but a factor of ten , or more , smaller than the large values predicted by the classical theory of pms evolution ( iben , hayashi , cameron , etc . ) that ignored the impact of protostellar evolution . another important feature of protostars is that their internal structure departs significantly from the assumption of thermal convection . in the standard pms theory , all stars were assumed to be fully convective objects as a result of the large initial radii . however , in low - mass protostars convection is due to central deuterium burning , while protostars more massive than about 2 m@xmath0 are radiatively stable in the inner regions , and possess a thick convective mantle maintained by deuterium burning . thus , these stars are _ thermally unrelaxed _ at the beginning of the pms evolution and must undergo _ non - homologous _ quasi - static contraction . these two properties have a profound impact on the calculation of pms evolutionary models and the resulting hr diagram . first , compared with the classical set of pms tracks , those shown in fig . [ hrd ] occupy a much more limited portion of the diagram . the reason is that the older tracks correspond to stellar radii too large to be attained during protostellar accretion . thus , the starting luminosities ( proportional to @xmath8 ) are much lower than previously envisioned . in addition , the surface temperature of these stars begin _ higher_. notice also that the birthline intersects the main sequence at a mass m@xmath9 m@xmath0 : this point represents the critical stellar configuration in which hydrogen burning has stopped gravitational contraction while the star is still growing in mass . _ more massive stars , therefore , have no pms phase at all , but appear directly on the main sequence once they are optically revealed_. the exact mass at which the birthline merges into the zams depends on the accretion history and increases as @xmath10 , as shown in fig . [ hrd ] . an important consequence is that observed pms stars should not be located to the right of the birthline . this prediction can be readily tested by the observations of large samples of t tauri and herbig ae / be stars in nearby star forming regions . the best studied young cluster in the solar neighborhood is the orion nebula cluster ( onc ) . its richness and high density make the onc an ideal target for the study of the stellar imf from 45 m@xmath0 to less than @xmath110.02 m@xmath0 ( e.g. , hillenbrand 1997 , slesnick et al . 2004 ) . the distribution of the optically visible stars in the hr diagram shown in fig . [ onchrd ] reveals several important features . first , the upper envelope of the stellar distribution is generally well matched by the birthline over the full mass spectrum . second , stars more massive than @xmath118 m@xmath0 are tightly grouped about the zams , while less massive ones diverge from it . third , the predicted endpoint of the birthline at m@xmath118 m@xmath0 also appears supported by the observational data : there are no massive stars still in the pms in spite of the fact that many low - mass stars have very young ages . since there is evidence that the most massive stars are also the youngest ( see sect . 5 ) , the lack of massive pms stars is quite significant . the onc is not an exception and similar results are obtained in the upper scorpius ob association , and in clusters such as ngc 6611 , ngc 3603 , and r136 in 30 doradus . the hr diagrams displayed in the right panel of fig . [ hrd ] summarize the current knowledge on the intermediate - mass population in several nearby ( @xmath12500 pc ) ob associations . again , it is clear that most stars are located between the birthline and the zams , with the most massive members already on the zams . ori b star ( discussed in sect . 5 ) is indicated.,width=264,height=264 ] the initial evolution of stars with mass @xmath52 m@xmath0 is the least affected by the protostellar conditions . the largest departure from the standard theory of pms stars concerns the history of deuterium burning since the latter can effectively start already during the accretion phase for masses above a few tenths of a solar mass . thus , t tauri stars , having consumed most of the available deuterium during the accretion phase , begin the contraction phase by simply descending along the vertical portion of the hayashi track . more relevant is the next episode in the life of a pms star , involving the nuclear burning and depletion of the second lightest isotope : lithium ( li ) as contraction proceeds , the critical temperature ( @xmath13 k ) for the reaction @xmath14li(p,@xmath15)@xmath16he is reached , and the initial li content is readily depleted in fully convective , sub - solar stars ( [email protected] m@xmath0 ) . objects with mass @xmath170.065 m@xmath0 never attain the ignition temperature and therefore maintain their initial abundance . the physics required to study the depletion history as a function of age has little uncertainty , since it depends mostly on the stellar mass and effective temperature ( @xcite ) . detailed numerical models show that fully convective stars in the range 0.50.2 m@xmath0 start to deplete li after about 510 myr , and completely destroy it in additional @xmath1110 myr ( e.g. , @xcite , @xcite ) . lower mass stars take much longer to burn li and , due to the strong temperature sensitivity of the energy generation rate , there is a sharp transition between fully depleted objects and those with the initial li content . thus , li - depletion can provide an independent clock to gauge stellar ages . from the hr diagram of the onc ( fig . [ onchrd ] ) , we see that the low - mass population has a large spread in luminosity ( @xmath1810 ) . does this reflect a corresponding spread in stellar ages ( factor of 10 or more ) , or is it merely a scatter due to the observational uncertainties ? this is an important question since the presence of older stars in young clusters bears directly on the issue of the duration of star formation . it is still unclear whether molecular clouds can sustain the production of stars for an extended period of time ( @xmath19 yr ) , longer than the typical free fall time ( @xmath20 yr ) ( e.g. @xcite , @xcite ) . therefore , the identification of an old population in the onc calls for an independent test , such as that provided by li - burning . figure [ lidepl ] shows the distribution in the hr diagram of a sample of onc members with mass in the range 0.41.0 m@xmath0 and isochronal ages greater than @xmath111 myr , together with the region of partial ( light shading ) and full ( dark shading ) depletion predicted by stellar evolution models . lithium abundances have been derived for all the stars in figure [ lidepl ] from measurements of the li i 670.8 nm doublet observed with flames on vlt ( @xcite ) . the main result is the discovery of significant depletion ( factor of 3 to 7 ) in four stars with estimated mass of @xmath110.4 m@xmath0 and age @xmath1110 myr . comparison with numerical and analytical models shows excellent agreement ( to within 10% ) between the isochronal age and lithium depletion time scale for two objects , the first case in lithium - poor pms stars . this is shown in figure [ lidepl_bild ] where the values of mass and age obtained from the hr diagram ( points with error bars ) are compared to the predictions of nuclear depletion following the semi - analytic method of bildsten et al . the derived ages of about 10 myr indicate that the onc does contain objects much older than the average age of the dominant population . thus , the star formation history of the onc extends long in the past , although at a reduced rate , in accordance with the empirical evidence found in the majority of nearby clusters and associations ( @xcite ) . _ even the onc did not come into existence in a single , rapid burst . _ ) values . ( from @xcite ) , width=264,height=264 ] let us now consider the intermediate - mass stars in the range 2 - 10 m@xmath0 . here , the impact of the protostellar initial consitions on the early evolution is the most profound and needs to be carefully checked . in the following , i will discuss two tests that allow to test the internal structure at the beginning of the pms phase and the predictions of the protostellar mass - radius relation . on their way to the zams , pms stars with mass @xmath21 cross the pulsation instability strip in the hr diagram . the instability is related to the classical @xmath22 and @xmath23 mechanisms in the h- and he - ionization regions near the stellar surface . in spite of the relatively short time spent in the instability strip ( about 0.050.1 of the kelvin - helmoltz time ) , more than 15 herbig ae stars have been identified as @xmath24 scuti - like pulsators with evidence for both linear and nonlinear modes ( see the review by @xcite ) . their position in the hr diagram is displayed in fig . [ stripcorot ] , along with the instability strip . the typical pulsation periods are in the range of several hours and amplitudes between hundredths and thousandths of magnitude . these studies are important since the relation between the pulsation period and the intrinsic stellar parameters ( l , t@xmath25 ) allow to obtain independent information on the evolutionary properties and in particular the stellar mass . additionally , at least in principle , asteroseismological techniques can be applied to infer the internal structure of the observed pulsators . in the present context , it is of particular interest the group of 4 stars with mass @xmath114 m@xmath0 located in the close vicinity of the birthline ( see fig . [ stripcorot ] ) . these stars represent an excellent sample to test the initial conditions for pms contraction . as shown in the upper panel of fig . [ figpuls ] , the track of a 4 m@xmath0 star with the protostellar initial conditions ( lower curve ) differs considerably from the standard one ( upper curve ) . the two curves join at a position near the red edge of the instability strip . the lower panel of fig . [ figpuls ] shows the internal luminosity profile of the 4 m@xmath0 star at the red edge : the solid and dashed lines are for the standard model and protostellar initial conditions , respectively . the latter displays a thermally unrelaxed luminosity profile while the former is characteristic of a fully radiative object . now , the pulsation characteristics ( periods , amplitudes ) are particularly sensitive to the internal conditions , especially in the subsurface regions where the perturbations are excited . detailed modeling indicate large differences in the excited linear and nonlinear modes ( @xcite ) . hopefully , these predictions can be tested by high sensitivity multisite monitoring campaigns and/or space observations ( such as the corot satellite scheduled for launch in 2006 ) . the upper curve is for the standard initial conditions , while the lower curve is for the protostellar conditions . _ bottom _ : the internal luminosity profile for the two models : the dashed curve is for the protostellar case . ( adapted from @xcite ) , width=264,height=264 ] the evolution of angular momentum in the pms phase shows that rotation is an important ingredient to understand young stars . both t tauri and herbig ae / be stars rotate at a small fraction ( @xmath261/10 ) of the breakup velocity , indicating that the angular momentum problem must be solved during the protostellar phase ( @xcite , @xcite ) . yet , the observed rotational velocities of pms stars are larger than those of more evolved stars of similar mass . thus , some angular momentum must be lost during pms contraction . although wind braking is certainly present already in the pms phase , it is not expected to dominate the rotational evolution since it operates on a time scale comparable or longer than @xmath27 ( @xcite ) . presently , the most popular mechanism for the removal of angular momentum involves magnetic star - disk interaction ( so - called , disk - locking hypothesis ; @xcite ; @xcite ; see also @xcite for a different view ) . according to this model , the stellar magnetic field threads the circumstellar disk and truncates it at a characteristic radius set by the balance between the pressure of the accreting gas and the magnetic pressure in the disk . accretion of disk material onto the star occurs along field lines , producing hot spots near the magnetic poles and locking the stellar rotation to the angular ( keplerian ) velocity of the disk at corotation . at the same time , magnetic torques transfer angular momentum away from the star to the disk . the location of the characteristic radius varies with the accretion rate : if @xmath28 increases , the radius decreases , thus increasing the rotation speed of the accreting gas and hence the star . also , a higher accretion rate affects the mass - radius relation producing larger ( proto)stellar radii . recently , @xcite et al . have used this argument to propose an interesting test on the initial conditions by measuring the projected rotational velocities in early - type ( b0b9 ) stars in the young clusters h and @xmath29 persei and comparing them with those of a sample of field stars of similar age ( @xmath1115 myr ) and mass ( 415 m@xmath0 ) . the basic idea is that if the accretion rates are higher in dense clusters , the rotational velocities of the h the @xmath29 persei cluster should be higher on average than the field population born in loose associations . indeed , strom et al . find that the late b - type ( 49 m@xmath0 ) cluster stars rotate faster ( factor of @xmath112 ) , while the difference is less for the early - type b stars . the higher rotation speed can be attributed to the enhanced spinup experienced during contraction by stars formed with a high accretion rate that produced large initial radii as a result of deuterium shell burning . on the other hand , for more massive stars , the difference in the initial radii are smaller due to their tendency to contract while accreting . if this evidence is corroborated by similar results in other young clusters , it would provide a new way to test the predictions of protostellar and pms models . what happens to massive stars ? how can we test mass and age from the location in the hr diagram ? how can we validate the accretion scenario developed for low- and intermediate - mass stars ? are very high accretion rates ( @xmath30 m@xmath0 yr@xmath2 ) ) really required for massive star formation , as repeatedly stated during the conference ? and , in case , is there evidence for an increase of @xmath31 with mass ? since most massive stars are in clusters , are they the first or the last to form ? returning to the case of the onc ( see fig . [ onchrd ] ) , we see that stars with mass @xmath328 m@xmath0 ( the trapezium stars ) are aligned along ( actually , slightly to the right of ) the zams . although all of them are part of binary / multiple systems , none of them is an eclipsing binary . thus , we have no direct access to their mass and evolutionary status . the only exception is the least massive of the trapezium stars , @xmath33 ori b ( see fig . [ onchrd ] ) , whose components ( each a binary system ) have a dynamical mass of m@xmath34=6.3 m@xmath0 and m@xmath35=2.5 m@xmath0 , respectively . when placed in the hr diagram , the secondary falls very close to the birthline and the two stars align on the same isochrone for @xmath36 yr , within the observational uncertainties ( @xcite ) . thus , this somewhat massive system is very young , confirming previous suggestions that massive stars tend to form relatively late in a cluster . unfortunately , current observations still lack the required resolution to derive the stellar parameters of the companions to the bright stars ( although individual masses can be derived in the next few years from orbit solutions ) . clearly , future observations should address this fundamental point . considering the formation mechanism , it is clear that the accretion scenario that works well for lower mass objects finds some natural problems at the highest masses . the main one is that radiation pressure begins to become significant in protostars of intermediate mass . however , infall can not be stopped as long as the accretion component dominates the total luminosity . we have seen that for @xmath37 m@xmath0 yr@xmath2 , a protostar joins the main sequence at a mass of @xmath118 m@xmath0 where it releases @xmath113000 l@xmath0 , less than the luminosity delivered by h - burning ( @xmath38 l@xmath0 ) at that mass . hence , the need for higher accretion rates ( for example increasing with mass , e.g. @xcite ) , indirectly suggested by the large mass _ loss _ rates inferred from the associated outflows . how to achieve the physical conditions for such high values of @xmath31 is still not at all clear , although centrally condensed turbulent cores may serve the purpose ( @xcite , but see dobbs & bonnell , this volume ) . perhaps , the early suggestion of an entirely different mode of formation based on growth by coalescence of dense molecular cores still represents a viable alternative ( e.g. @xcite ) . interestingly , evidence for possible interactions between cores at high infall rates has been found in the case of the dense clumps of ngc 2264c ( @xcite ) . we have reached an important conclusion : in order to distinguish between the two plausible explanations , massive protostars vs. young zams stars , the critical quantity to determine observationally is @xmath31 . future high angular resolution millimeter observations will reveal whether bright embedded objects are characterized by _ high _ values of the accretion rate ( protostars ) or not ( zams stars ) . given the large distances of massive star forming regions ( @xmath321 kpc ) , one has to wait for the completion of ( sub)mm - wave arrays such as alma to reach the desired spatial resolutions . owing to the extremely high extinction along the line of sight , the direct observation of the source ( photosphere or inner accretion disk ) responsible for the high infrared luminosity of the orion bn / kl nebula is precluded . however , the source can be indirectly seen through the scattered light of the surrounding bipolar nebula ( @xcite ) . morino et al . obtained a 2 @xmath39 m spectrum of the reflection nebula tracing the radiation that escaped the dust surrounding the massive protostar possibly along the polar axis of a circumstellar disk / torus . the spectrum shows absorption features due to co ro - vibration bands formed at t@xmath114500 k. if the observed 2 @xmath39 m radiation is emitted by the protostellar photosphere , the luminosity estimated from the reflected light and the effective temperature yield a protostellar radius larger than @xmath11300 r@xmath0 . with such large radii , the required mass accretion rate onto the protostar should exceed @xmath40 m@xmath0 yr@xmath2 . alternatively , the inner portion of an accretion disk around a massive protostar could be responsible for the observed near - infrared radiation . for a steady accretion disk , the k - band radiation with an effective temperature of 4500 k would be emitted by an annular region at a distance of @xmath11200 r@xmath0 , assuming a central protostar of 10 m@xmath0 . given the size of the emitting region , accretion rates of @xmath41 m@xmath0 yr@xmath2 are required in order to produce to produce the observed k - band luminosity . both possibilities , a huge protostar or an extremely active disk , represent extreme situations with respect to standard models of protostellar evolution . however , the difference between these pictures is large enough to be tested observationally . in particular , the co bandhead _ absorption _ lines at 1.6 and 2.3 @xmath39 m are expected to have distinctively different profiles in the two cases if observed at sufficiently high spectral resolution ( @xmath3210000 ) . this diagnostic differs from the study of co _ emission _ profiles observed towards a number of distant massive young stars ( see contribution by blum ) where the emission lines can not be photospheric in origin , but trace the circumstellar material close to the central object , possibly in keplerian rotation . in the case of the ir reflection nebula , the _ absorption _ co features could have an origin in the protostellar photosphere and the velocity resolved profile can test this hypothesis on this and similar objects . the main points of this review are : * protostellar evolution via accretion greatly affects the properties of young pms stars , particularly in the intermediate- and high - mass regime . * for low - mass pms stars , li - depletion during early pms evolution can be used to gauge the age and age spread in young clusters , such as the onc . * the peculiar initial conditions of pms stars with mass @xmath1148 m@xmath0 ( herbig ae / be ) can be tested using pulsations and rotation as reliable diagnostics . * the disk accretion scenario predicts that stars with mass @xmath32815 m@xmath0 should be on the zams once optically visible . currently , there is no evidence for massive pms stars . * the empirical agreement of the position of the birthline in several clusters / associations indicates that @xmath42 m@xmath0 yr@xmath2 can account for the formation of stars and their attendant disks up to @xmath111020 m@xmath0 . * for more massive protostars , the magnitude of @xmath43 is still unknown . co bandhead _ absorption _ lines can be used to infer the accretion rate around massive protostars , the most elusive quantity to measure .
young stars on their way to the zams evolve in significantly different ways depending on their mass . while the theoretical and observational properties of low- and intermediate - mass stars are rather well understood and/or empirically tested , the situation for massive stars ( 1015 m@xmath0 ) is , to say the least , still elusive . on theoretical grounds , the pms evolution of these objects should be extremely short , or nonexistent at all . observationally , despite a great deal of effort , the simple ( or bold ) predictions of simplified models of massive star formation / evolution have proved more difficult to be checked . after a brief review of the theoretical expectations , i will highlight some critical test on young stars of various masses .
quantum clock synchronization protocols are of fundamental interest in quantum information , since they can illustrate how information about time is encoded in quantum systems . in general , there are presently two approaches to the problem . the first is based on the correlations between photon arrival times , or the related arrival times of optical signals detected by homodyne detection @xcite . the second approach is based on the internal time evolution of quantum systems @xcite . although the latter approach requires an effective suppression of decoherence and is therefore much more challenging to implement , it might be of greater fundamental interest , since it allows a very general treatment of time in quantum mechanics . specifically , it can show how the time evolution of quantum systems affects the non - classical correlations between entangled quantum systems . initially , it was shown that two - party quantum clock synchronization protocols can be used for efficient clock synchronization by using the enhanced sensitivity of bipartite entangled states to small time differences between the measurements performed by the two parties @xcite . later , krco and paul @xcite extended this idea to a multi party version , where a w - state was used to simultaneously provide bipartite entanglement between a central clock and several other parties . however , the bipartite entanglement obtained from the w - state decreases rapidly with an increase in the number of clocks . ben - av and exman @xcite pointed out that this is a weakness of the w - state that can be overcome by using other dicke states instead . specifically , they showed that the optimal bipartite entanglement for this kind of protocol is obtained by using the symmetric dicke states , where half of the qubits are in the 0 state and half are in the 1 state . interestingly , none of these protocols uses the specific properties of multipartite entanglement . the reason for this may be that it is a straightforward matter to measure and evaluate the correlations between two parties , while genuine multi - partite entanglement is characterized by more complicated correlations that involve the measurement results observed at all locations . here , we consider the question of whether this different type of entanglement could be used for clock synchronization by constructing a protocol that accesses the maximal entanglement of ghz - type states through an appropriate combination of measurement and communication between the parties . the result should be significant both for determining the limits of multi party clock synchronization and for our general understanding of time in multi - partite entanglement . as we discuss below , a protocol using ghz - type states requires a specific division of the parties into groups during each distribution of the entangled qubits , since there is no ghz - type state that is both an energy eigenstate and symmetric in all parties . the logic of the two party protocol can then be applied to the collective information of all measurements in the two groups . by selecting an appropriate set of divisions , it is thus possible to use the complete entanglement of the ghz - state to efficiently synchronize all n clocks simultaneously . in the following , we introduce the protocol , evaluate its efficiency and compare it with the efficiencies obtained from bipartite entanglement in the protocols based on parallel two party synchronization @xcite and on the use of symmetric dicke states @xcite . here , we consider a type of clock synchronization protocol where entangled qubits are distributed to various parties holding the individual clocks @xcite . each quantum system is described as a two - level spin precessing around the z - axis at a fixed frequency @xmath0 . for experimental realizations , it would be important to keep decoherence effects to a minimum , e.g. by using nuclear spins precessing in the intrinsic field of a molecule or crystal @xcite . however , it should be kept in mind that the present research is not motivated by such technical considerations , but by an interest in the fundamental nature of quantum clocks , as introduced in the semininal work by buzek et al . @xcite . in this spirit , we assume that after state preparation , the evolution of the internal spin state only depends on the passage of time . the problem of clock synchronization can then be reduced to the problem of identifying the time differences between time - dependent measurements performed on the different clocks . since clock synchronization should not depend on a knowledge of the time needed for state distribution , the multipartite entangled states used should be energy eigenstates . it is therefore not possible to use ghz states that are superpositions of the two extremal eigenstates of energy , where all qubits are in the same state of their local energy basis . to obtain an energy eigenstate without changing the multipartite entanglement , half of the local energy eigenstates should be flipped by appropriate local unitary transformations . if the qubits are arranged so that the first half of the qubits is unflipped and the second half of the qubits is flipped , this @xmath1-partite entangled energy eigenstate can be given in the energy basis as @xmath2 here and in the following , we assume an even number of parties @xmath1 . the states @xmath3 and @xmath4 are local energy eigenstates with energies @xmath5 and @xmath6 , respectively . the state given by eq.([maximal entanglement ] ) divides the qubits into two groups . to ensure clock synchronization between all parties , it is necessary that no two parties are always members of the same group . this is achieved by distributing the qubits in different ways , so that each party sometimes receives a qubit from the unflipped group , and sometimes receives a qubit from the flipped group . to describe each distribution , we define a sequence @xmath7 , where @xmath8 . if the qubit of the @xmath9 clock owner is a flipped qubit , @xmath10 , if not , @xmath11 . since the numbers of flipped and unflipped qubits are equal , the number of possible distributions @xmath12 is given by the binomial coefficient @xmath13 . in the most simple version of the protocol , the division into groups can be decided randomly in each run , with equal probabilities for each distribution @xmath12 . after the distribution of the qubits to the locations of the different clocks , each of the parties measures a time dependent observable @xmath14 on its qubit when their local clock points to a specific time . the observable measured at a time @xmath15 can be written as @xmath16 the eigenvalues of the measurement outcomes are @xmath17 . the eigenstates corresponding to the measurement outcomes are equal superpositions of @xmath18 and @xmath19 , where the phase now depends on the time at which the measurement is performed . as a result , this measurement achieves the maximal time sensitivity for local qubit measurements . the time sensitivity of the maximal multi - partite entanglement of the ghz - type energy eigenstate given in eq.([maximal entanglement ] ) originates from the coherence between the components @xmath20 and @xmath21 . this coherence , which represents the full multi - partite entanglement of the state , changes the probability of the collective measurement outcome depending on the product of the coherences between the @xmath18 and @xmath19 components in the eigenstates representing the local measurement outcomes . as a result , the time sensitivity of multi - partite entanglement can be represented by the espectation value for the product of all outcomes , @xmath22 . if the actual measurement times of the parties are given by @xmath23 the expectation value of this product is @xmath24 since the local measurements represent the maximal time sensitivity for the local qubits , and since the coherence that characterizes multi - parite entanglement of the ghz - type can only be observed in the product of all local measurement outcomes , we can conclude that the time dependence shown in eq.([average number ] ) is the strongest dependence on the measurement outcomes @xmath25 that can be achieved with ghz - type states and local measurements . a protocol that makes use of the time dependent correlations between all of the @xmath26 measurements therefore accesses the full power of maximal multipartite entanglement for clock synchronization . to access the time sensitivity of the ghz - type state , all parties must share their measurement results and determine the product of all outcomes . effectively , the @xmath1 parties cooperate to measure a single @xmath1-particle interference fringe that is sensitive to the collective phase given by @xmath0 times the difference between all measurement times of the unflipped qubits ( @xmath27 ) and the measurement times of all flipped qubits ( @xmath28 ) . significantly , the use of maximal multi - partite entanglement for clock synchronization critically depends on simultaneous access to all measurement outcomes . it therefore requires classical communication between all the parties , as indicated in fig . the full sensitivity of maximal multipartite entanglement only becomes available for use in the clock synchronization process , if all of the parties cooperate . [ ht ] , scaledwidth=20.0% ] in the final step of the synchronization protocol , the clock owners have to estimate the time differences between their respective local clocks and a standard time . since the present protocol is fully cooperative , with complete symmetry between all parties , it seems natural to define this standard time as the average of all clock times . in the following , we therefore discuss the synchronization to average time . to achieve synchronization to a standard clock , one can simply change the adjustments of the time , so that instead of changing his or her own time , the owner of the standard clock makes everyone else subtract his or her time adjustment from theirs . to ensure that all parties are treated equally , it is possible to use a random distribution of qubits , so that every distribution @xmath29 of flipped and unflipped qubits is equally likely . to keep track of the different distributions , we assign an index @xmath30 to each , so that the elements of each sequence are given by @xmath31 . the total time difference that defines the phase shift in the multipartite interference fringe observed in the @xmath32 measurement of the distribution with index @xmath30 is then @xmath33 the time differences @xmath34 can be estimated from the outcome statistics of the measurements with an accuracy of @xmath35 , where @xmath36 is the number of times that the distribution @xmath30 is received and measured . after a sufficiently large number of measurements , all parties have the same estimates for all possible time differences @xmath34 . however , the implications of each @xmath34 are different for each party . specifically , each party @xmath37 can obtain the difference of times @xmath34 with @xmath38 and the times @xmath34 with @xmath39 , @xmath40 for @xmath41 , the coefficient in the sum is always @xmath42 , so that the time @xmath25 of the local clock always enters into the sum with a positive value . since all the other times enter into the sum equally , and since the number of @xmath42 coefficients and @xmath43 coefficients is exactly equal , the result can be expressed in terms of the difference between the time @xmath25 and the average @xmath44 of all times other than @xmath25 , @xmath45 the average of all times @xmath46 is obtained by the weighted average of @xmath25 and @xmath47 times @xmath48 . hence , the difference between the local time @xmath25 and the average time @xmath46 can be given by , @xmath49 after this value is determined by each party , it can be subtracted from each local clock time @xmath25 to adjust the clock times so that they correspond to the average time @xmath46 . to determine the efficiency of a clock synchronization protocol , it is necessary to evaluate the precision with which the parties can estimate the adjustment time @xmath50 . in general , this precision is limited by the statistical variance of the measurement results . as mentioned above , the estimation errors for the time differences @xmath34 are given by @xmath51 , where @xmath36 is the number of times that the distribution @xmath30 was measured . since the adjustment times @xmath50 are linear functions of the @xmath34 , it is sufficient to find the sum of the quadratic errors with the appropriate coefficients to obtain the adjustment errors @xmath52 if each distribution @xmath30 is measured an equal number of times , @xmath36 can be expressed as the total number of measurements @xmath53 divided by the number of possible distributions @xmath13 . likewise , the sum over @xmath30 reduces to a simple multiplication with the number of possibilities . in the end , the estimation error for each adjustment time is given by @xmath54 in the limit of large @xmath1 , this error is simply @xmath55 , independent of the number of parties participating in the clock synchronization . this means that the maximally multipartite entangled states can be used to synchronize @xmath1 clocks in parallel , without any loss of precision when additional parties are added . since the parallel synchronization of @xmath1 clocks can also be achieved by performing separate synchronizations of @xmath47 clocks with the same standard clock , it is not immediately clear whether multipartite entanglement has any advantages over multiple bipartite entangled states . in the following , we therefore analyze the efficiency of multipartite clock synchronization using the initial proposal for quantum clock synchronization between two parties @xcite . the complete multi party protocol is illustrated in fig.2 . there are @xmath1 spatially separated unsynchronized clocks , one of which is the standard clock . the remaining @xmath47 clock owners synchronize their clocks with this standard clock using bipartite entanglement and classical communication . for this purpose , the owner of the central clock must share @xmath47 maximally entangled two qubit states with all of the other parties for each measurement . effectively , each step of the protocol uses a @xmath56 qubit state given by @xmath57 here , every second qubit is held by the owner of the central clock . at a predetermined time @xmath15 , the owner of the central clock measures the value of @xmath58 on all of her @xmath47 qubits @xmath37 , where @xmath37 is the index of the party that holds the qubit entangled with the qubit @xmath59 . likewise , the other parties measure the value of @xmath60 on their individual qubits according to their local times @xmath25 . the central clock then communicates each result of @xmath58 to the party concerned . after a sufficiently large number of measurements @xmath53 , the owner of clock @xmath37 can then determine the expectation value of the product , @xmath61 the clock owners can then determine @xmath62 directly and adjust their clocks accordingly . [ ht ] , scaledwidth=20.0% ] the efficiency of clock synchronization can be evaluated by considering the estimation error @xmath63 for each estimate of adjustment time @xmath62 . for @xmath53 measurements , this error is given by @xmath64 . thus the precision of the time estimates in this protocol is exactly equal to the result for the protocol using multipartite entanglement . however , the parallel distribution of bipartite entanglement requires @xmath56 qubits for each measurement , as compared to only @xmath1 qubits for the multipartite entangled protocol . in terms of the required number of qubits , the use of multipartite entangled states can thus increase the efficiency by a factor of two . effectively , the main effect of multipartite entanglement seems to be that the need for multiple reference qubits held by the owner of the central clock is removed by allowing the parties to use the @xmath47 qubits of all the other parties as a collective reference instead . previous multi party clock synchronization protocols were based on parallel clock synchronization using the bipartite entanglement available from w - states @xcite or from symmetric dicke states @xcite . in particular , ben - av and exman showed that the symmetric dicke state is optimal in the sense that it maximizes the bipartite entanglement between the single qubit held by the owner of the central clock and the qubits held by all of the other parties @xcite . their protocol is illustrated in fig . it uses the same measurement and communication procedure as in the parallel distribution of entanglement , but with only a single qubit at the central clock for a total number of @xmath1 qubits per measurement - the same number as our ghz state protocol , and almost half the number of qubits used in the parallel distribution protocol . [ ht ] , scaledwidth=20.0% ] the symmetric dicke states is an equal superposition of all energy eigenstates with half of the qubits in the @xmath18 state and half in the @xmath19 state , @xmath65 the bosonic symmetry of the state means that the qubits tend to be found in the same superposition states of @xmath18 and @xmath66 , resulting in positive correlations between the values of @xmath67 obtained by the different parties at the same time @xmath25 . specifically , the correlation between the measurement at the central clock and the measurement at clock @xmath37 is given by @xmath68 at the maximal time derivative of the expectation value , the error in the adjustment time @xmath69 for @xmath53 measurements is given by @xmath70 . in the limit of large @xmath1 , this error is equal to twice the error of our ghz state protocol and the parallel distribution protocol . hence , this protocol requires four times as many qubits to achieve the same accuracy as the ghz state protocol , and twice as many qubits as the parallel distribution protocol . the reduction in qubit number over parallel distribution of bipartite states is therefore more than offset by the loss of sensitivity in each individual measurement due to the reduction in the available bipartite entanglement . we can now summarize our results in terms of the accuracy of clock synchronization achieved with a given number of qubits . since the timescale is defined by the resonant frequency @xmath0 of the qubit dynamics , it is convenient to define the relative accuracy as @xmath71 . for the ghz - type multipartite entanglement , the accuracy of @xmath53 measurements using @xmath72 qubits is then given by @xmath73 for high @xmath1 , the accuracy is equal to the number of qubits per party , so the accuracy of the multi party protocol scales linearly with the ratio of qubits and parties , @xmath74 . significantly , the straightforward extension of the bipartite protocol by parallel distribution of entangled qubit pairs performs only half as well . specifically , the accuracy of @xmath53 measurements using @xmath75 qubits is @xmath76 the need for extra reference qubits held by the owner of the central clock therefore rapidly reduces the efficiency of each qubit to half the value achieved by the protocol using maximal multipartite entanglement . finally , the protocol using the simultaneous bipartite entanglement between a single central qubit and @xmath47 others achieves a sensitivity reduced by a factor of @xmath77 due to the reduction in bipartite entanglement associated with the increase in entangled partners for each qubit . the accuracy of @xmath53 measurements using @xmath78 qubits is therefore @xmath79 in the limit of high @xmath1 , this is a reduction to one quarter of the ghz - type protocol , twice as much as the reduction in accuracy due to the additional reference qubits in the parallel distribution protocol . we have shown how the maximal @xmath1-partite entanglement of ghz - type stated can be used for multi party clock synchronization by randomly dividing the parties into two groups during each run and sharing the measurement results with all other parties to determine the adjustments necessary to set each local clock to the average time of all clocks . the accuracy of clock synchronization corresponds to the accuracy achieved by @xmath47 bipartite protocols in parallel , but the number of qubits used is reduced by half . oppositely , the previously proposed use of symmetric dicke states uses the same number of qubits , but the accuracy is only one quarter due to the reduced amount of bipartite entanglement . our results show that the full power of maximal multipartite entanglement can improve the performance of clock synchronization by reducing the number of qubits needed to achieve a given accuracy by a factor of two when compared to the most efficient use of bipartite entanglement . although this is clearly an improvement , it is much less than the improvements of sensitivity when a single parameter is estimated using multi - partite entangled probes . the reason for this limited improvement is that @xmath1 different clock times must be estimated from the same measurement result , leading to a reduction of precision that exactly compensates the gain caused by the increased sensitivity to an average shift in time . from the viewpoint of an individual clock owner , multi - partite entanglement merely replaces the single central clock used in parallel clock synchoronization with bipartite entangled states with the collective of all other @xmath47 clock owners . overall , the efficiency increases by a factor of two , because the simultaneous role of all clock owners as participants and as reference overcomes the need for additional reference qubits . in this sense , multi - partite entanglement simply represents the simultaneous availability of quantum correlations to all parties , without any increase to the individual time sensitivities . from a practical viewpoint , the use of multi - patite entanglement may be difficult , since the loss of a single qubit will completely destroy the essential quantum coherence of the state . to obtain results close to the ones described here , the probability of local losses or dephasing errors must be kept far below @xmath80 . oppositely , this sensitivity to decoherence also highlights the cooperative nature of the protocol : if only a single party sends wrong information , the synchronization of the clocks becomes impossible . thus , clock synchronization with maximal multipartite entanglement also highlights the cooperative nature of multi - party quantum protocols . in conclusion , the analysis presented here shows how the full power of maximal multipartite entanglement can be used to improve the performance of clock synchronization if all of the parties involved cooperate to share their measurement information . the results may provide interesting insights , both into the role of entanglement in clock synchronization protocols , and into the fundamental nature of time - dependent quantum correlations . part of this work has been supported by the grant - in - aid program of the japanese society for the promotion of science , jsps .
we propose a multi party quantum clock synchronization protocol that makes optimal use of the maximal multipartite entanglement of ghz - type states . to realize the protocol , different versions of maximally entangled eigenstates of collective energy are generated by local transformations that distinguish between different groupings of the parties . the maximal sensitivity of the entangled states to time differences between the local clocks can then be accessed if all parties share the results of their local time dependent measurements . the efficiency of the protocol is evaluated in terms of the statistical errors in the estimation of time differences and the performance of the protocol is compared to alternative protocols previously proposed .
we are envisioning a smart iot system , addressing the key challenges as described below : the first challenging problem is that devices are not interoperable at any level with each other since most of the time technologies differ from one to another . for instance , in contemporary iot applications multiple competing application level protocols such as constrained application protocol ( coap ) , message queue telemetry transport ( mqtt ) and extensible messaging and presence protocol ( xmpp ) are becoming popular @xcite . each protocol possesses unique characteristics and messaging architecture helpful for different types of iot applications . however , a smart iot application architecture should be independent of messaging protocol standards , while also providing integration and translation between various popular messaging protocols . similarly to proprietary protocols , at the data level , devices do not use common terms or vocabulary to describe interoperable iot data . the traditional paradigm of the iot service model provides unformated data names as `` raw '' sensor data . this `` raw '' sensor data does not contain any aggregated description ( usually representation through semantic annotations ) and requires specialized knowledge and manual effort in order to build practical applications . much of the current use of iot is targeted to a single domain and most of the times the number of sensors are duplicated unnecessarily . for instance , temperature sensors in a building primarily used for a heating , ventilating , and air conditioning ( hvac ) application . however , values produced by temperature sensors could be used in other applications such as fire detection . the primary advantage of using common sensors into various applications is that it can reduce development , maintenance and deployment costs and promote device reusability . to enable crossdomain applications and address interoperability issues , a smart iot system is needed to publish their outputs and to describe device information in a wellunderstood format with added metadata and machineprocessable formats , thus making devices accessible and usable in multiple applications . in iot systems , users are primarily interested in realworld entities ( such as people , places and things ) and their highlevel states ( e.g. , deriving snowfall from temperature and precipitation measurements ) rather than raw output data produced by sensors attached with these entities . to achieve this requirement , a smart iot system has to provide highlevel knowledge that can map sensors to realworld entities and output of raw sensor to highlevel states @xcite . to easily develop iot applications at a large scale with little or no human intervention , a smart iot system should leverage semantic web properties , and follow standards . the web of knowledge also plays a relevant role , by defining the rules and mechanisms to associate information in order to produce knowledge . see `` smart iot : iot as a human agent , human extension , and human complement '' in @xcite for the first definition of smart iot , which highlights the challenge of interoperating and integrating the data and information . we are envisioning a smart iot system that enables good decision making and actions . figure [ fig : architecture ] shows an architecture overview of the system inspired by @xcite . the architecture largely divided into three layers by their functions : an example of a smart iot application is represented below in figure [ fig : app ] taken from the first column in this series ( from data to decisions and actions : climbing the data , information , knowledge , and wisdom ( dikw ) ladder @xcite ) . the lowest level shows `` 150 '' which is a blood pressure reading ( sensor / device data ) . the next level shows labeled ( semantically annotated ) data or information . the third level represents knowledge that is based on the latest nih guidance used by clinicians , this information represents a medical condition of `` elevated blood pressure '' . and yet this knowledge is not actionable the clinician needs to decide whether this is due to hyperthyroidism or hypertension , which is needed before a proper medication can be prescribed . in today s internet of things landscape the cyber , virtual and physical worlds are largely disconnected , requiring a lot of manual efforts to integrate , find , and use information in a meaningful way . to realize the application as we discussed above , we are envisioning a smart iot system that enables good decision making and actions . figure [ fig : architecture ] shows an architecture overview of the system inspired by @xcite . the architecture largely divided into three layers by their functions : 1 . * accessing things ( physical ) * : this layer is responsible for turning a device such that an application can interact with it . the gateways use devicespecific protocols to retrieve data produced by resourceconstrained devices . the gateways add semantics to data to unify them , by using semantic web languages ( such as rdf , rdfs , owl ) and domain ontologies . * deducing new knowledge ( virtualization ) * : the second layer is dedicated to frameworks managing unified data available in standard formats produced by the physical layer . it mainly infers high level knowledge using reasoning engines performed on data and by exploiting the web of knowledge available online . such enriched data is provided to the cyber layer to build smart systems , applications and services . * composing services ( cyber ) * : this layer facilitates developers so that they can build largescale and meaningful iot applications on top of the virtualization layer . the goal of this level is to drastically reduce the iot application development , thus enabling rapid prototyping and encourage interoperability of services . to date , to the best of our knowledge , no concrete and robust technical approaches have been designed to build semantic interoperability for iot yet . recently , some semantic interoperability approaches applied to iot are being designed @xcite . an iot stack to ensure interoperability has been designed in @xcite . to define such architecture , semantic interoperability should be provided as explained in @xcite where it is introduced the idea of an effective approach to bring together metadata , information modeling abstractions and ontologies , as well as the model application domain . hereafter , we describe the above three layers , mapping into an example with details that assist developers to design and build smart iot applications . the first layer is responsible for turning a device into such that an application can interact with it . the most straightforward way of accessing a device is to expose it and its services directly through apis . this is applied when a device can support http / web service ( ws ) and tcp / ip and can host an http server . however , integrating resourceconstrained devices into the internet is difficult because internet protocols such as http , tcp / ip are too complex and resourcedemanding . to achieve the integration , typically a gateway node is required . to provide interoperability , we can implement necessary technologies at the resource sufficient gateway node . desai et al . have proposed the concept of `` semantic gateway as service '' ( sgs),shown in figure [ fig : gateway ] , that can act as a bridge between resourceconstrained devices and iot application services @xcite . here , the iot application services typically collect data from the various gateway nodes and provide user or event specific services using graphics interfaces , notifications or applications . the sgs architecture broadly provides three functionalities : first , it connects external sink device to the gateway component that supports different protocols such as mqtt , xmpp or coap . second , externally the gateway interfaces cloud services or other sgss via different protocols such as rest or publish / subscribe . third , it annotes data acquired from the sink nodes using w3c semantic sensor networks ( ssn ) @xcite and domainspecific ontologies before forwarding data to the gateway interface service . the main benefit is that semantic annotation of sensor data by utilizing a standard mechanism and vocabulary can provide interoperability between iot vertical silos . semantic web community has created and optimized standard ontologies for sensor observation , description , discovery and services . by integrating these annotated data and providing semantic web enabled messaging interface , a third party service can convert heterogeneous sensor observations to higher level abstractions . a gateway device such as sgs , discussed above , separates physical level implementation of device to iot application services . it provides endpoint to iot application services using a resource interface via rest and publish / subscribe mechanism . the semantic annotation of the sensor data obtained from the gateway assists the iot services to implement analysis and reasoning algorithms , described in the next layer . the second layer is dedicated to frameworks managing data and deducing new knowledge . in iot , most of the time , raw data is just a number ( e.g. , 38 ) . humans implicitly know that data is associated to a specific unit ( e.g. , degree celsius ) and a specific sensor ( e.g. , thermometer ) . smart iot systems need to interconnect data produced by various sensors to understand the meaning of data to automatically take decisions or to provide suggestions . dealing with interoperability of heterogeneous data is required to build smarter iot systems . data is stored in different files ( e.g. , csv , excel ) and structured with different models ( e.g. , ontology , schemas ) . to deal with heterogeneous data , semantic web technologies bring several benefits : ( 1 ) unify data , ( 2 ) link iot data to external knowledge bases , ( 3 ) explicitly add metadata ( i.e. , semantic enrichment / enhancement ) , and ( 4 ) deduce new knowledge . semantic web technologies enable interconnecting knowledge graphs . by interconnecting iot data with such knowledge graphs ( datasets and ontologies used to structure the datasets ) , iot systems are becoming smarter . one current form of knowledge graphs that is widely available and useful is `` linked open data '' ( lod ) @xcite . major internet companies such as facebook and google are buildding private knowledge graphs based on schema.org ( agreement on common schemas to structure data ) that is widely adopted by search engines , in addition to components of lod such as dbpedia , wikipedia and/or wikidata . such knowledge graphs are build to get access to the information requested more easily and in a automatic way . companies have chosen to use different representation for semantic data , that include unlabeled and labeled graphs , w3c ratified semantic web languages such as rdf and rdfs to explicitly describe the data , and owl , a language to describe ontologies / vocabularies . such semantic technologies provide a basis to later infer high level abstractions from sensor data . connecting unified semanticsenriched iot data to the knowledge bases available on the web has a huge potential to build smart systems . for instance , by connecting health measurements to healthcare knowledge bases enable interpreting the raw data itself . for instance , from a body temperature data and by reusing knowledge bases on the web , fever symptom can be deduced . interconnecting better such knowledge bases , particularly in smart iot is required . currently , open ontology catalogues and dataset catalogues are not interconnecting with each other . new systems are required to link ontologies and datasets , along with the methods to deduce new knowledge from structured data . taking inspiration from `` linked open data '' and `` linked open vocabularies '' ( lov ) @xcite initiatives , `` linked open reasoning '' should be designed and aligned to such initiatives . different processes and steps are required for combining data from heterogeneous sources and for building innovative and interoperable applications . figure [ fig : seg ] introduces the seg 3.0 methodology that seeks to meet these requirements and comprises the following steps : ( 1 ) composing , ( 2 ) modeling , ( 3 ) linking , ( 4 ) reasoning , ( 5 ) querying , ( 6 ) services , and ( 7 ) composition of services @xcite . the seg 3.0 methodology encourages the vision to enhance semantic interoperability from data to enduser applications , which is inspired from the sharing and reusing based approach as depicted in figure [ fig : seg].the seg 3.0 methodology comprises : * * linked open data ( lod ) * is an approach to share and reuse data . previous work regarding ` linked sensor data ' @xcite do not provide any tools for visualizing or navigating through iot datasets . for this reason , we envision the design of linked open data cloud for internet of things ( cloudiot ) to share , browse and reuse data produced by sensors . * * linked open vocabularies ( lov ) * is an approach to share and reuse the models / vocabularies / ontologies [ vandenbussche et al . 2015 ] . to ensure reusability and high quality ontologies , lov did not reference any ontologies when they do not follow the best practices . due to this requirement , almost all ontologies for iot and relevant domain ontologies were not referenced by lov since iot community does not yet know the best practices . to overcome this limitation , the linked open vocabularies for internet of things ( lov4iot ) has been designed , a dataset of almost 300 ontologybased iot projects referencing and classifying : ( 1 ) iot applicative domains , ( 2 ) sensors used , ( 3 ) ontology status ( e.g. , shared online , best practices followed ) , ( 4 ) reasoning used to infer high level knowledge , and ( 5 ) research articles related to the project . this dataset contains a background knowledge required to add value to the data produced by devices . * * linked open reasoning ( lor ) * is an innovative approach to share and reuse the way to interpret data to deduce new information ( e.g. , machine learning algorithm used , reusing rules already designed by domain experts ) . sensorbased linked open rules ( slor ) is a dataset of interoperable rules ( e.g. , if then else rules ) used to interpret data produced by sensor data @xcite . such rules are executed with an inference engine which updates the triple store with additional triples . for example , the execution of the rule `` if the body temperature is greater than 38 degree celsius than fever '' updates the triple store with the high level knowledge fever. slor is inspired from the idea of linked rules which provides a language to interchange semantic rules but not the idea of reusing existing rules . * * linked open services ( los ) * is an approach to share and reuse the services / applications @xcite @xcite . composition of services is required to build complex applications . services can be implemented according to restful principles or with the help of semantic web technologies to enhance interoperability ( e.g. , owls ) . this approach could be extended for designing a set of interoperable services . sharing and reusing data is insufficient . the entire chain from linked open data ( lod ) to linked open services ( los ) should be shared and reused to enhance interoperability and get meaningful knowledge from data . having this vision in mind , the models , the reasoning and the services associated to the data would be interoperable with each other . this entire chain , called seg 3.0 methodology , has been implemented within the m3 framework @xcite @xcite and extended within the fiesta iot eu platform . m3 enables fast prototyping of iot applications using semantic web technologies to semantically annotate sensor data , deduce new knowledge and combine iot applicative domains . m3 is a semantic engine mainly focused on data interoperability and could be used in other eu projects such as fiestaiot , openiot and vital . fiestaiot aims to achieve interoperability of data , testbeds and experiments by using semantic web technologies . one of the component of the fiestaiot project , called `` experiment as a service '' demonstrates the proof of concept of the `` linked open services '' approach . figure [ fig : seg ] highlights an endtoend scenario from raw value ( e.g. , 38 ) to the final application ( e.g. , naturopathy to suggest home remedies when fever is detected ) . such applications can be developed through the use of the m3 framework . developers can build largescale and meaningful iot applications and services on top of devices and iot data . the goal of this level is to drastically reduce the iot application development @xcite , thus enabling rapid prototyping . this layer also gets closers to enduser ( or domain experts with a limited programming expertise ) and enables them to create intelligent applications on top of smart things . in the following , we describe application development approaches for building iot applications . * generalpurpose programming . * currently , development of iot is performed at the nodelevel , by experts of embedded and distributed systems , who are directly concerned with operations of each individual device @xcite . for example , developers use generalpurpose programming languages ( such as javascript , c , java , android , python ) and target a particular middleware api or nodelevel service to communicate data . the key advantage of this approach is that it allows the development of efficient systems based on the complete control over individual devices . however , it is unwieldy for iot applications due to the heterogeneity of systems . * macroprogramming . * it provide abstractions to specify highlevel collaborative behaviors , while hiding lowlevel details such as message passing or state maintenance from stakeholders . a classic example of macroprogramming is nodered . it is a visual tool for wiring together hardware devices , apis , and online services . it provides browserbased environment for creating eventdriven applications , bridging between physical and cyber services . it contains nodes that can be dragged and dropped into an editor . each node offers different functionality that can range from a simple debug functionality to accessing sensors via gateways ( e.g. , raspberry pi ) . macroprogramming is viable approach compared to generalpurpose programming . however , this approach largely lacks proper software development methodology ( e.g. , modular design , separation of concerns ) . that results into a difficult to reuse and platformdependent design . * cloudbased platforms . * to improve development effort , the cloudbased platforms provide apis that provide functions to implement common functionality of iot applications such as sending and storing data to cloud for data visualization . moreover , these platforms provide textual and visual programming running on the cloud to write a custom application logic . they provide abstractions to specify highlevel collaborative behaviours while hiding lowlevel details such as message passing . an example of cloudbased approach is ibminternet of things foundations ] . it is a fully managed and cloudhosted service that makes it simple to derive value from physical devices . using abstractions ( it is called as recipes ) , developers can connect devices to the internet , send sensing data securely to the cloud using the open and lightweight mqtt messaging protocol . from there , developer can leverage various cloudbased services such as dashboard services to visualize and derive insight from the collected data , storage services to store data for historical purpose . cloudbased platform is a viable approach compared to the generalpurpose programming languages . it reduces development efforts by providing cloudbased apis to implement common functionality . second advantage is that because application logic is centrally located , this approach offers the ease deployment and evolution . however , this approach sacrifices direct nodetonode communication . this characteristics restricts developers interms of functionality such as innetwork aggregation or direct nodetonode communication locally . third , application logic largely runs on a central cloud , thus an application relies on the availability of cloud provider . so , it may not be suitable for some critical applications . * model driven development ( mdd ) . * macroprogramming and cloudbased approach reduce the application development effort . however , they lead to a platformdependent design . to address this issue , mdd approaches have been proposed . it applies the basic separation of concerns principle both vertically and horizontally . vertical separation principle reduces the application development complexity by separating the specification platform independent model ( pim ) of the system functionality from its platform platform specific model ( psm ) such as programming languages . horizontal separation principle reduces the development complexity by describing a system using different system views , each view describing a certain facet of the system . an example of mdd approach is iotsuite @xcite @xcite that aims to make iot application development easy for developers . it achieves this aim by integrating a set of highlevel languages to specify an iot application . it provides automation techniques to parse the specifications written using these highlevel languages and generate platformspecific code . the iotsuite integrates three highlevel languages that abstract platformspecific complexity ( i.e. , horizontal separation of concerns ) : ( 1 ) domain language to describe domainspecific features of an iot application , ( 2 ) architecture language to describe applicationspecific functionality of an iot application , ( 3 ) deployment language to describe deploymentspecific features consisting information about a physical environment where devices are deployed . the iotsuite is supported by automation techniques such as codegenerator that generates platformspecific code by parsing the specification written using the supported highlevel programming languages ( i.e. , vertical separation of concern ) . ensuring semantic interoperability within iot is really challenging since physical , virtual and cyber layers deal with heterogeneity of hardware devices , protocols , data and reasoning mechanisms to infer high level knowledge . the seg 3.0 methodology explained above is mainly focused on data interoperability and is a first step towards building interoperable enduser iot applications and services . as discussed in the first column in this series ( see previous issue ) , iot deployment is experimenting a fast adoption because of the foreseen positive impact to change all the aspects of our lives . this is accompanied by corresponding variety or heterogeneity for all aspects of iot ecosystem , including data , communication and application development frameworks . the vision of smart iot , also discussed in the previous column , envisages hiding this heterogeneity and corresponding complexity , while enabling development of applications based on intelligent real - time processing of data produced variety of sensors , along with relevant knowledge . semantic methods and semantic web standards are key enablers of the three requisite layers of a smart iot system : accessing things ( iots ) , understanding iot data and deducing new knowledge using background / domain knowledge and structured data , and developing composing services . this article ends with a review of four approaches to application development in a smart iot ecosystem . s. chauhan , p. patel , f. c. delicato , and s. chaudhary . a development framework for programming cyber - physical systems . in _ proceedings of the 2nd international workshop on software engineering for smart cyber - physical systems _ , sescps 16 , pages 4753 , new york , ny , usa , 2016 . s. chauhan , p. patel , a. sureka , f. c. delicato , and s. chaudhary . demonstration abstract : iotsuite - a framework to design , implement , and deploy iot applications . in _ 2016 15th acm / ieee international conference on information processing in sensor networks ( ipsn ) _ , pages 12 , april 2016 . m. compton , p. barnaghi , l. bermudez , r. garca - castro , o. corcho , s. cox , j. graybeal , m. hauswirth , c. henson , a. herzog , et al . the ssn ontology of the w3c semantic sensor network incubator group . , 17:2532 , 2012 . a. gyrard . a machine - to - machine architecture to merge semantic sensor measurements . in _ proceedings of the 22nd international conference on world wide web _ , www 13 companion , pages 371376 , new york , ny , usa , 2013 . a. gyrard , c. bonnet , and k. boudaoud . helping iot application developers with sensor - based linked open rules . in _ proceedings of the seventh international workshop on semantic sensor networks in conjunction with the 13th international semantic web conference ( iswc14 ) , riva del garda , italy _ , pages 1923 , 2014 . a. gyrard , c. bonnet , k. boudaoud , and m. serrano . assisting iot projects and developers in designing interoperable semantic web of things applications . in _ 2015 ieee international conference on data science and data intensive systems _ , pages 659666 , dec 2015 . a. gyrard , s. k. datta , c. bonnet , and k. boudaoud . cross - domain internet of things application development : m3 framework and evaluation . in _ future internet of things and cloud ( ficloud ) , 2015 3rd international conference on _ , pages 916 . ieee , 2015 . a. gyrard and m. serrano . connected smart cities : interoperability with seg 3.0 for the internet of things . in _ 2016 30th international conference on advanced information networking and applications workshops ( waina ) _ , pages 796802 , march 2016 . d. pfisterer , k. romer , d. bimschas , o. kleine , r. mietz , c. truong , h. hasemann , a. kroller , m. pagel , m. hauswirth , m. karnstedt , m. leggieri , a. passant , and r. richardson . spitfire : toward a semantic web of things . , 49(11):4048 , november 2011 . m. serrano , h. n. m. quoc , m. hauswirth , w. wang , p. barnaghi , and p. cousin . open services for iot cloud applications in the future internet . in _ world of wireless , mobile and multimedia networks ( wowmom ) , 2013 ieee 14th international symposium and workshops on a _ , pages 16 . ieee , 2013 . m. serrano , h. n. m. quoc , d. le phuoc , m. hauswirth , j. soldatos , n. kefalakis , p. p. jayaraman , and a. zaslavsky . defining the stack for service delivery models and interoperability in the internet of things : a practical case with openiot - vdk . , 33(4):676689 , 2015 . d. soukaras , p. patel , h. song , and s. chaudhary . . in _ the 4th international workshop on computing and networking for internet of things ( comnet - iot ) , co - located with 16th international conference on distributed computing and networking ( icdcn ) _ , page 6 , 2015 .
the internet of things ( iot ) is experiencing fast adoption in the society , from industrial to home applications . the number of deployed sensors and connected devices to the internet is changing our perspective and the way we understand the world . the development and generation of iot applications is just starting and they will modify our physical and virtual lives , from how we control remotely appliances at home to how we deal with insurance companies in order to start insurance schemes via smart cards . this massive deployment of iot devices represents a tremendous economic impact and at the same time offers multiple opportunities . however , the potential of iot is underexploited and day by day this gap between devices and useful applications is getting bigger . additionally , the physical and cyber worlds are largely disconnected , requiring a lot of manual efforts to integrate , find , and use information in a meaningful way . to build a connection between the physical and the virtual , we need a knowledge framework that allow bilateral understandings , devices producing data , information systems managing the data and applications transforming information into meaningful knowledge . the first column in this series in the previous issue of this magazine titled `` internet of things to smart iot through semantic , cognitive , and perceptual computing , '' reviews iot growth and potential that have energized research and technology development , centered on aspects of artificial intelligence to build future intelligent system . this column steps back and demonstrates the benefits of using semantic web technologies to get meaningful knowledge from sensor data to design smart systems .
the aim of this project is : * to determine column densities of simple diatomic + molecules toward new 80 targets in our database * to check relations between abundances of ch , ch@xmath0 , cn molecules and e(b - v ) basing on statistically meaningful sample of objects * to find out the new relation between column densities of ch and h@xmath1 molecules based on new results concerning ch molecule ( abundances of molecular hydrogen are published by savage et al . 1977 , rachford et al . 2002 ) . our observing material consists of spectra of 168 stars containing cn and/or ch and ch@xmath0 features . it was acquired using five echelle spectrometers : * maestro fed by the 2-m telescope of the observatory at peak terskol ( see http://www.terskol.com /telescopes/3-camera.htm ) * feros spectrograph , fed with the 2.2 m eso telescope in chile + ( see http://www.ls.eso.org/lasilla/sciops/2p2/e2p2m/feros/ ) * fiber - fed echelle spectrograph installed at 1.8-m telescope of the bohyunsan optical astronomy observatory ( boao ) in south korea * harps spectrometer , fed with the 3.6 m eso telescope in chile + ( see http://www.ls.eso.org/lasilla/sciops/3p6/harps/ ) * uves spectrograph at paranal in chile . for more information see : + http://www.eso.org/sci/facilities/paranal/instruments/uves to obtain column densities of diatomic molecules we used : * in the case of cn features with ew@xmath214m ( cn features are saturated ) we used the method of roth & meyer ( 1995 ) . * when ch line of a - x system was saturated we used the sum of column densities obtained from unsaturated 3886 and 3890 lines of b - x system . * in the case of saturated ch@xmath0 a - x ( 0 , 0 ) band at 4232 we used unsaturated ( 1 , 0 ) band centered near 3957 . oscillator strengths in the case of each transition are given in table 1 . * the interstellar reddening correlates better with ch than ch@xmath0 molecule . a very poor relation is observed in the case of cn . * certain objects ( hd 27778 , 147889 , 154368 , 161056 , 204827 , 208501 ) demonstrate high abundances of diatomic molecules in relation to e(b - v ) + ch&cn abundant clouds * hd 34078 is untypical object in case of which ch and ch@xmath0 molecules are overabundant in relation to e(b - v ) ch&ch@xmath0 abundant cloud * the relation between column densities of ch and h@xmath1 molecules is very tight ( see also results of danks , federman and lambert 1984 , weselak et al . 2004 ) . new relation with correlation coefficient equal to 0.95 is obtained on the basis of 53 stars . authors acknowledge the financial support : jk and tw acknowledge that of the polish state during the period 2007 - 2010 ( grant n203 012 32/1550 ) . we are grateful to y. byaletski . , g. locurto , g.a . galazutdiov and f.a . musaev for their help in ackquiring the spectra used and many helpful suggestions . danks a.c . , federman s.r . , lambert d.l . , 1984 , a&a , 130 , 62 rachford b.l . , snow , t.p . , tumlinson , j. , et al . , 2002 , apj , 577 , 221 roth , k.c , meyer , d.m . , 1995 , apj , 441 , 129 savage , b.d . , bohlin , r.c . , drake , j.f . , budich , w. , 1977 , apj , 216 , 291 van dishoeck , e.f . , black , j.h . , 1986 , apjs , 62 , 109 van dishoeck , e.f . , black , j.h . , 1989 , apj , 340 , 273 weselak , t. , galazutdinov , g.a . , musaev , f.a . , kreowski , j. , 2004 , a&a , 414 , 949 weselak , t. , galazutdinov , g.a . , musaev , f.a . , kreowski , j. , 2008a , a&a , 479 , 149 weselak , t. , galazutdinov , g.a . , musaev , f.a . , kreowski , j. , 2008b , a&a , 484 , 381
we analyze mutual relations between interstellar extinction e(b - v ) and column densities of diatomic molecules such as cn , ch and ch cation observed along translucent sightlines . observational material acquired using five echelle spectrographs , situated in both northern and southern hemispheres , was used in this study . while column densities of ch , ch@xmath0 correlate reasonably tightly with e(b - v ) the similar relation built for cn shows a large scatter . this result suggests that ch and ch@xmath0 molecules are quite well spatially correlated to interstellar dust while cn requires another conditions to be formed . the ch molecule is closely related to h@xmath1 as it was demonstrated already ; the fact is interesting since h@xmath1 is believed to be formed on grain surfaces rather than in the gas phase .
pulsar radio emission beam has been widely attempted to interpret in terms of emission in purely dipolar magnetic field . gangadhara & gupta ( 2001 ) have estimated the emission heights of different radio pulse components in psr b0329 + 54 based on the aberration retardation phase shift , and the revised estimates are given by dyks , rudak , & harding ( 2003 ) . here we solve the viewing geometry and estimate the altitudes from which observer can receive the radio waves . consider a magnetic dipole situated at the origin with magnetic axis ( @xmath2 ) inclined by @xmath3 with respect to the rotation axis @xmath4 and rotated by @xmath5 around @xmath6 let @xmath7 be the line of sight , where @xmath8 and @xmath9 is the line of sight impact parameter . in a relativistic flow , the emitted radiation is beamed in the direction of field line tangent @xmath10 , so at any instant the observed radiation comes from a spot in the magnetosphere where the tangent vector points in the direction @xmath11 of observer . for receiving such radiation the semi opening angle of emission beam @xmath12 must be approximately equal to the opening angle of field lines @xmath13 therefore , the magnetic _ colatitude _ @xmath14 is given by @xmath15 next , the magnetic _ azimuth _ @xmath16 of the emission point can be obtained by finding @xmath17 which is parallel to @xmath18 @xmath19 for @xmath20 on leading side the maximum value for @xmath21 allowed by the viewing geometry is @xmath22 which in turn allows to find the maximum pulse window @xmath23 where @xmath5 is the pulse phase at which @xmath21 approaches @xmath24 using @xmath25 and @xmath26 for psr b0329 + 54 , we find @xmath27 @xmath28 and @xmath29 pulsar radio emission is generally believed to be coherent curvature radiation by secondary pair plasma streaming along the dipolar magnetic field lines . the curvature emission peaks at the characteristic frequency ( e.g. , eq . 45 , ruderman & sutherland 1975 ) . for a given frequency and a lorentz factor @xmath30 we can estimate the radius of curvature @xmath31 , which in tern allows to find the field line constant . so , using @xmath32 and 390 we estimated the emission height ( see , fig . 1a ) of radiation at 325 mhz and 606 mhz , respectively . on the other hand by accepting the emission heights derived from the aberration - retardation phase shift , we estimated @xmath1 and @xmath31 expected : @xmath33 @xmath34 for 325 mhz emission , and @xmath35 @xmath36 for 606 mhz . the polar cap with foot location of emission associated field lines is given in fig . 1b , where @xmath37axis is chosen to be parallel to @xmath38 and x to lie in the @xmath39-@xmath38 plane . it is nearly elliptical with radius of 164 m and 171 m in x and y directions .
the beamed emission by relativistic sources moving along the magnetic dipolar field lines occur in the direction of tangents to the field lines . to receive such a beamed radiation line - of - sight must align with the tangent within the beaming angle @xmath0 where @xmath1 is the particle lorentz factor . by solving the viewing geometry , in an inclined and rotating dipole magnetic field , we show that at any given pulse phase observer can receive the radiation only from the specific altitudes . we find the outer conal emission is received from the higher altitudes than the inner conal components including the core . at any pulse phase , low frequency emission comes from the higher altitudes than the high frequency emission . as an application of our model , we have applied it to explain the emission heights of conal components in psr b0329 + 54 . epsf.sty 1.25 in 0.125 in 0.25 in
such as the nonlinear schrdinger equation ( nlse ) , the complex cubic - quintic ginzburg - landau equation ( ccqgle ) is one of the most intensively studied equation describing weakly nonlinear phenomena in dissipative systems @xcite . thus , much work has been done about the features and the properties of this equation as well as its numerous applications such as nonlinear waves , superconductivity , superfluidity @xcite , bose - einstein condensation , bnard convection @xcite and nonlinear optics @xcite . + solitary waves ( or solitons ) are self localized solutions of certain nonlinear pdes describing the evolution of a dissipative system @xcite . whereas traditional solitons are stationary in time and preserve their shape upon interaction , some dissipative soliton solutions of the ccqgle are non stationary . in hamiltonian systems , stationary solitons exist as a result of a balance between diffraction / dispersion and nonlinearity . however , a dissipative system is known to be far from equilibrium and is defined by energy exchanges with external sources . thus , there is no conserved quantity , which implies that they are not hamiltonian systems . in dissipative systems , solitons are obtained by balancing gain and loss . as a result , whereas solutions are defined by one parameter family in hamiltonian systems , in dissipative systems solutions are obtained with varying amplitude and width that are fixed by the parameters of the equation . therefore , the inverse scattering method @xcite which is used to calculate solutions of integrable and some non integrable hamiltonian systems can not be used in our case . in this paper , we present new classes of soliton solutions of the ccqgle and we analyze both their shape and energy evolution that depends strongly on the parameters used , and choice of initial conditions . the development of powerful numerical tools over the last decades lead us to use an accurate spectral method for the spatial discretization , together with a @xmath0 order adams - bashforth scheme for temporal discretization . due to the intensive computations codes in fortran 90 , we run simulations on zeus cluster ( 256 nodes dual xeon 3.2 ghz processors , 1024 kb cache 4 gb with myrinet ) at embry - riddle aeronautical university . for each class of solitons , simulations lasted from several hours to several days . numerical simulations are based on the 2d ccqgle model with transverse laplacian @xmath1 that takes into account spatial diffraction in the paraxial wave approximation . quintic terms are essential in order to obtain stable solitons @xcite @xmath2 @xmath3 is the normalized envelope of the field which is a complex function , @xmath4 and @xmath5 are the transverse spatial coordinates and @xmath6 is the propagation distance along the optical fiber ( or cavity depth ) . the physical parameters of the system are : @xmath7 linear gain / loss , @xmath8 angular spectral filtering , @xmath9 second - order diffraction coefficient , @xmath10 nonlinear gain / loss , @xmath11 nonlinear dispersion , @xmath12 saturation of the nonlinear gain / loss , and @xmath13 saturation of the nonlinear refractive index . the solutions to will be computed using a fourier spectral method in which we will monitor the energy @xmath14 . for a localized solution , @xmath15 is finite and changes smoothly while the solution stays within the region of existence of the soliton . when @xmath15 changes abruptly there is a bifurcation and the solution jumps from a branch of solitons that become unstable to another branch of stable solitons , or vice versa . as soon as the solution becomes unstable , @xmath15 diverges until infinity or collapses to @xmath16 . for a certain class of solutions , @xmath15 will evolve periodically in some regime , and will converge to a finite value @xcite . the main goal of this study is to find continuous parameter ranges around a fixed set in which each class of solitons remains stable . first , an initial set of parameters published in @xcite is used for the elliptic , filament , pulsating and exploding classes , while for creeping we use parameters published in @xcite , see table 1 . starting from these values , for each class we vary one parameter in a range of width @xmath17 while keeping all the other parameters constant . when we are outside of the range of existence of the solitons , we return to the initial value , and vary the other remaining parameters one by one . after several simulations , we noticed that @xmath10 and @xmath7 seemed to be the more sensitive , i.e. , small changes of @xmath10 or @xmath7 implied significant changes of the solution . thus , a step of @xmath18 has been chosen for these parameters whereas a step of @xmath19 has been used for @xmath8 , @xmath12 and @xmath13 . however , this behavior is not a general rule , so , in some cases , a smaller step has been used for the last three parameters . .initial sets of parameters for 2d solitons . [ cols="^,^,^,^,^,^,^,^",options="header " , ] in conclusion , we have found fascinating ( 2 + 1)d structures in dissipative media predicted by the complex cubic - quintic ginzburg - landau equation . different types of solutions , spinning or not , such as elliptic vortex , filament , pulsating , exploding and creeping solitons have been categorized by focusing on their energy , shape and phase evolution . regions of existence of these types of soliton have also been calculated in the five dimensional parameter space given by the physical parameters of the equation . the study revealed interesting ranges and bifurcations for each structure . whereas non spinning structures have mainly lead to stable stationary solitons , due to bifurcations spinning structures , symmetric or asymmetric , split and have revealed to be no spin solitons . by changing the initial conditions , and increasing the vorticity , allowed us to observe for the first time new types of solutions such as disorganized exploding and propeller solitons . akhmediev n. , soto - crespo j. and town g. , _ pulsating solitons , chaotic solitons , period doubling , and pulse coexistence in mode - locked lasers : complex ginzburg - landau equation approach _ , physical review e * 63 * ( 2001 ) 56602 . soto - crespo j. , akhmediev n. , devine n. and meja - corts c. , _ transformations of continuously self - focusing and continuously self - defocusing dissipative solitons _ , optics express * 16 * ( 2008 ) 15388 - 15401
by using spectral methods , we present seven classes of stable and unstable structures that occur in a dissipative media . for different parameters of the equation , we find ranges of existence of stable structures ( spinning elliptic , pulsating , stationary , organized exploding ) , and unstable structures ( filament , disorganized exploding , creeping ) . by varying initial conditions , and vorticity , we find a reacher behavior of solutions in the form of creeping - vortex ( propellers ) , spinning rings and spinning `` bean - shape '' solitons . each class differentiates from the other by distinctive features of their shape and energy evolution , as well as domain of existence .
a measurement of @xmath33 opens an important window on the non - perturbative strong physics involving heavy quarks . the basic framework of the theory is well understood , however , there is still much speculation - predictions for the width range from @xmath34 to @xmath35 @xcite . the level splitting in the @xmath36 sector is not large enough to allow real strong transitions . therefore , a measurement of the width of the @xmath0 gives unique information about the strong coupling constant in heavy - light meson systems . the total width of the @xmath0 is the sum of the partial widths of the strong decays @xmath3 and @xmath37 and the electromagnetic decay @xmath38 . we can write the width in terms of strong couplings , @xmath39 and @xmath40 , and an electromagnetic coupling , @xmath41 : @xmath42 where the momenta are those for the indicated particle in the @xmath0 rest frame , and @xmath43 is the fine structure constant . this can be rewritten using the isospin relationship @xmath44 and relating @xmath45 to a universal strong coupling between heavy vector and pseudoscaler mesons to the pion , @xmath46 , with @xmath47 where @xmath48 is the pion decay constant . all this yields @xmath49 the width of the @xmath0 only depends on @xmath46 @xcite since the contribution of the electromagnetic decay with branching fraction @xmath50% @xcite can be neglected . the measurement of @xmath46 is needed in the extraction of @xmath51 in semileptonic @xmath52 decays @xcite . prior to this measurement , the @xmath0 width was limited to be less than @xmath53 at the @xmath54 confidence level by the accmor collaboration @xcite . the limit was based on 110 signal events reconstructed in two @xmath55 decay channels with a background of 15% . this contribution describes a measurement of the @xmath0 width with the cleo ii.v detector where the signal , in excess of 11,000 events , is reconstructed through a single , well - measured sequence , @xmath56 , @xmath57 . consideration of charge conjugated modes are implied throughout this paper . the level of background under the signal is less than 3% in our loosest selection . the challenge of measuring the width of the @xmath0 is understanding the tracking system response function since the experimental resolution exceeds the width we are trying to measure . candidates with tracks that have mismeasured hits , errors in pattern recognition , and large angle coloumb scattering are particularly dangerous because the signal shape they project is broad and the errors for these events can be underestimated , resulting in events that can easily influence the parameters of a breit - wigner fitting shape . we generically term such effects `` tracking mishaps . '' a difficulty is that there is no physical calibration for this measurement . the ideal calibration mode would have a large cross - section , a width of zero , decay with a rather small energy release to three charged particles one of which has a much softer momentum distribution than the other two which decay through a nearly zero width resonance with a measurable flight distance . such a mode would allow us to disentangle detector effects from the underlying width but no such mode exists . therefore , to measure the width of the @xmath0 we depend on exhaustive comparisons between a geant @xcite based detector simulation and our data . we addressed the problem by selecting samples of candidate @xmath0 decays using three strategies . first we produced the largest sample from data and simulation by imposing only basic tracking consistency requirements . we call this the _ nominal _ sample . second we refine the nominal sample selecting candidates with the best measured tracks by making very tight cuts on tracking parameters . there is special emphasis on choosing those tracks that are well measured in our silicon vertex detector . this reduces our nominal sample by a factor of thirty and , according to our simulation , has negligible contribution from tracking mishaps . we call this the _ tracking selected _ sample . a third alternative is to select our data on specific kinematic properties of the @xmath0 decay that minimize the dependence of the width of the @xmath0 on detector mismeasurements . the nominal sample size is reduced by a factor of three and a half and , again according to our simulation , the effect of tracking problems is reduced to negligible levels . we call this the _ kinematic selected _ sample . in all three samples the width is extracted with an unbinned maximum likelihood fit to the energy release distribution and compared with the simulation s generated value to determine a bias which is then applied to the data . these three different approaches yield consistent values for the width of the @xmath0 giving us confidence that our simulation accurately models our data . the cleo detector has been described in detail elsewhere . all of the data used in this analysis are taken with the detector in its ii.v configuration @xcite . this work mainly depends on the tracking system of the detector which consists of a three - layer , double sided silicon strip detector , an intermediate ten - layer drift chamber , and a large 51-layer helium - propane drift chamber . all three are in an axial magnetic field of 1.5 tesla provided by a superconducting solenoid that contains the tracking region . the charged tracks are fit using a kalman filter technique that takes into account energy loss as the tracks pass through the material of the beam pipe and detector @xcite . the data were taken in symmetric @xmath1 collisions at a center of mass energy around 10 gev with an integrated luminosity of 9.0/fb provided by the cornell electron - positron storage ring ( cesr ) . the nominal sample follows the selection of @xmath58 candidates used in our @xmath59 mixing analysis@xcite . our reconstruction method takes advantage of the small cesr beam spot and the kinematics and topology of the @xmath60 decay chain . the @xmath61 and @xmath62 are required to form a common vertex . the resultant @xmath55 candidate momentum vector is then projected back to the cesr luminous region to determine the @xmath55 production point . the cesr luminous region has a gaussian width @xmath63 m vertically and @xmath64 m horizontally . it is well determined by an independent method@xcite . this procedure determines an accurate @xmath55 production point for @xmath55 s moving out of the horizontal plane ; @xmath55 s moving within 0.3 radians of the horizontal plane are not considered . then the @xmath65 track is refit constraining its trajectory to intersect the @xmath55 production point . this improves the resolution on the energy release , @xmath66 , by more than 30% over simply forming the appropriate invariant masses of the tracks . the improvement to resolution is essential to our measurement of the width of the @xmath0 . our resolution is shown in figure [ fig : errorcompare ] and is typically 150 kev . the good agreement between monte carlo and data demonstrates that the kinematics and sources of uncertainties on the tracks , such as the number of hits used and the effects of multiple scattering in detector material , are well modeled . to further improve the quality of reconstruction in our nominal sample , we apply some kinematic cuts to remove a small amount of misreconstructed signal and background . figure [ fig : ppisoft ] shows the distribution of the momentum [ cols="^,^ " , ] [ tab : systematic ] uncertainties on the width of the @xmath0 and @xmath67 . we have measured the width of the @xmath0 by studying the distribution of the energy release in @xmath3 followed by @xmath68 decay . we have done this in three separate samples , one that is minimally selected , a second that reduces poorly measured tracks due to misassociated hits and non - gaussian scatters in the detector material , and a third that takes advantage of the kinematics of the decay chain to reduce dependence on mismeasurements of kinematic parameters . the resolution on the energy release is well modeled by our simulation , with agreement between the sources of the resolution as predicted by the simulation and analytic calculations . the largest sources of uncertainty are imperfect modeling of the dependence of the mean energy release on the kinematics of the decay chain , the simulation of the error on the energy release , and correlations among the parameters of the fit to the energy release distribution . with our estimate of the systematic uncertainties for each of the three samples being essentially the same we chose to report the result for the sample with the smallest statistical uncertainty , the minimally selected sample , and obtain @xmath69 where the first uncertainty is statistical and the second is systematic . we note that if we form an average value taking into account the statistical correlations among our three measures we get a result that is nearly identical with equation [ eq : result ] since the average is dominated by the result with the smallest statistical uncertainty . this is the first measurement of the width of the @xmath0 , and our measurement corresponds to a strong coupling@xcite @xmath70 and @xmath71 this is consistent with theoretical predictions based on hqet and relativistic quark models , but higher than predictions based on qcd sum rules . we also measure the mean value for the energy release in @xmath3 decay @xmath72 where the first error is statistical and second is systematic . combining this with the mass of the charged pion , 139.570 mev , with an uncertainty less than 1 kev @xcite , we calculate @xmath73 this agrees with the value from the particle data group , @xmath74 mev , from a global fit of all flavors of @xmath75@xmath76 mass differences . it also agrees well with the best previous measure from a single experiment that includes an evaluation of systematic uncertainties from accmor at @xmath77 mev @xcite . we thank d. becirevic , i. i. bigi , g. burdman , a. khodjamirian , p. singer , and a. l. yaouanc for valuable discussions . we gratefully acknowledge the effort of the cesr staff in providing us with excellent luminosity and running conditions . m. selen thanks the pff program of the nsf and the research corporation , and a.h . mahmood thanks the texas advanced research program . this work was supported by the national science foundation , the u.s . department of energy , and the natural sciences and engineering research council of canada . v.m.belyaev _ et al . _ phys . rev . d 51 , 6177 ( 1995 ) contains a recent survey summarizing and referencing previous theoretical work . p. singer acta phys . b30 3849 ( 1999 ) , j. l. goity and w. roberts jlab - thy-00 - 45 ( hep - ph/0012314 ) , k. o. e. henriksson , _ et al . phys . a 686 , 355 ( 2001 ) , and m. di pierro and e. eichten hep - ph/0104208 appear since that survey . m. wise , phys . d * 45 * , r2188 ( 1992 ) , g. burdman and j. f. donoghue , phys . b * 280 * , 287 ( 1992 ) , and t. yan _ et al . _ , phys . d * 46 * , 1148 ( 1992 ) [ erratum - ibid . d * 55 * , 5851 ( 1992 ) ] . d. becirevic and a. l. yaouanc , jhep * 9903 * , 021 ( 1999 ) contains a recent review referencing previous theoretical work . ( cleo collaboration ) , phys . * 80 * , 3919 , ( 1998 ) . g. burdman , z. ligeti , m. neubert and y. nir , phys . rev . d * 49 * , 2331 ( 1994 ) . et al . _ , * b 278 * , 480 ( 1992 ) . r. brun _ et al . _ , geant3 users guide , cern dd / ee/84 - 1 . y. kubota _ et al . _ , ( cleo collaboration ) , nucl . instrum . methods phys . res . , sect . a * 320 * , 66 ( 1992 ) ; t. hill , nucl . methods phys . res . , sect . a * 418 * , 32 ( 1998 ) . p. billior , nucl . methods phys . res . , sect . a * 225 * , 352 ( 1984 ) . r. godang _ ( cleo collaboration ) , phys . lett . * 84*,5038 ( 2000 ) d. cinabro _ et al . _ , clns 00/1706 , physics/0011075 . d. e. groom _ et al . _ ( particle data group ) , eur.phys.j . * c 15 * , 1 ( 2000 ) .
we present the first measurement of the @xmath0 width using 9/fb of @xmath1 data collected near the @xmath2 resonance by the cleo ii.v detector . our method uses advanced tracking techniques and a reconstruction method that takes advantage of the small vertical size of the cesr beam spot to measure the energy release distribution from the @xmath3 decay . we find @xmath4 kev . we also measure the energy release in the decay and compute @xmath5 mev / c@xmath6 . 6.5 in 9.0 in -0.50 in 0.00 in 0.00 in a. anastassov,@xmath7 e. eckhart,@xmath7 k. k. gan,@xmath7 c. gwon,@xmath7 t. hart,@xmath7 k. honscheid,@xmath7 d. hufnagel,@xmath7 h. kagan,@xmath7 r. kass,@xmath7 t. k. pedlar,@xmath7 j. b. thayer,@xmath7 e. von toerne,@xmath7 m. m. zoeller,@xmath7 s. j. richichi,@xmath8 h. severini,@xmath8 p. skubic,@xmath8 a. undrus,@xmath8 v. savinov,@xmath9 s. chen,@xmath10 j. w. hinson,@xmath10 j. lee,@xmath10 d. h. miller,@xmath10 e. i. shibata,@xmath10 i. p. j. shipsey,@xmath10 v. pavlunin,@xmath10 d. cronin - hennessy,@xmath11 a.l . lyon,@xmath11 w. park,@xmath11 e. h. thorndike,@xmath11 t. e. coan,@xmath12 y. s. gao,@xmath12 y. maravin,@xmath12 i. narsky,@xmath12 r. stroynowski,@xmath12 j. ye,@xmath12 t. wlodek,@xmath12 m. artuso,@xmath13 k. benslama,@xmath13 c. boulahouache,@xmath13 k. bukin,@xmath13 e. dambasuren,@xmath13 g. majumder,@xmath13 r. mountain,@xmath13 t. skwarnicki,@xmath13 s. stone,@xmath13 j.c . wang,@xmath13 a. wolf,@xmath13 s. kopp,@xmath14 m. kostin,@xmath14 a. h. mahmood,@xmath15 s. e. csorna,@xmath16 i. danko,@xmath16 v. jain,@xmath17 k. w. mclean,@xmath16 z. xu,@xmath16 r. godang,@xmath18 g. bonvicini,@xmath19 d. cinabro,@xmath19 m. dubrovin,@xmath19 s. mcgee,@xmath19 a. bornheim,@xmath20 e. lipeles,@xmath20 s. p. pappas,@xmath20 a. shapiro,@xmath20 w. m. sun,@xmath20 a. j. weinstein,@xmath20 d. e. jaffe,@xmath21 r. mahapatra,@xmath21 g. masek,@xmath21 h. p. paar,@xmath21 a. eppich,@xmath22 t. s. hill,@xmath22 r. j. morrison,@xmath22 h. n. nelson,@xmath22 r. a. briere,@xmath23 g. p. chen,@xmath23 t. ferguson,@xmath23 h. vogel,@xmath23 j. p. alexander,@xmath24 c. bebek,@xmath24 b. e. berger,@xmath24 k. berkelman,@xmath24 f. blanc,@xmath24 v. boisvert,@xmath24 d. g. cassel,@xmath24 p. s. drell,@xmath24 j. e. duboscq,@xmath24 k. m. ecklund,@xmath24 r. ehrlich,@xmath24 p. gaidarev,@xmath24 l. gibbons,@xmath24 b. gittelman,@xmath24 s. w. gray,@xmath24 d. l. hartill,@xmath24 b. k. heltsley,@xmath24 l. hsu,@xmath24 c. d. jones,@xmath24 j. kandaswamy,@xmath24 d. l. kreinick,@xmath24 m. lohner,@xmath24 a. magerkurth,@xmath24 h. mahlke - krger,@xmath24 t. o. meyer,@xmath24 n. b. mistry,@xmath24 e. nordberg,@xmath24 m. palmer,@xmath24 j. r. patterson,@xmath24 d. peterson,@xmath24 d. riley,@xmath24 a. romano,@xmath24 h. schwarthoff,@xmath24 j. g. thayer,@xmath24 d. urner,@xmath24 b. valant - spaight,@xmath24 g. viehhauser,@xmath24 a. warburton,@xmath24 p. avery,@xmath25 c. prescott,@xmath25 a. i. rubiera,@xmath25 h. stoeck,@xmath25 j. yelton,@xmath25 g. brandenburg,@xmath26 a. ershov,@xmath26 d. y .- j . kim,@xmath26 r. wilson,@xmath26 b. i. eisenstein,@xmath27 j. ernst,@xmath27 g. e. gladding,@xmath27 g. d. gollin,@xmath27 r. m. hans,@xmath27 e. johnson,@xmath27 i. karliner,@xmath27 m. a. marsh,@xmath27 c. plager,@xmath27 c. sedlack,@xmath27 m. selen,@xmath27 j. j. thaler,@xmath27 j. williams,@xmath27 k. w. edwards,@xmath28 a. j. sadoff,@xmath29 r. ammar,@xmath30 a. bean,@xmath30 d. besson,@xmath30 x. zhao,@xmath30 s. anderson,@xmath31 v. v. frolov,@xmath31 y. kubota,@xmath31 s. j. lee,@xmath31 r. poling,@xmath31 a. smith,@xmath31 c. j. stepaniak,@xmath31 j. urheim,@xmath31 s. ahmed,@xmath32 m. s. alam,@xmath32 s. b. athar,@xmath32 l. jian,@xmath32 l. ling,@xmath32 m. saleem,@xmath32 s. timm,@xmath32 and f. wappler@xmath32 @xmath7ohio state university , columbus , ohio 43210 + @xmath8university of oklahoma , norman , oklahoma 73019 + @xmath9university of pittsburgh , pittsburgh , pennsylvania 15260 + @xmath10purdue university , west lafayette , indiana 47907 + @xmath11university of rochester , rochester , new york 14627 + @xmath12southern methodist university , dallas , texas 75275 + @xmath13syracuse university , syracuse , new york 13244 + @xmath14university of texas , austin , texas 78712 + @xmath15university of texas - pan american , edinburg , texas 78539 + @xmath16vanderbilt university , nashville , tennessee 37235 + @xmath18virginia polytechnic institute and state university , blacksburg , virginia 24061 + @xmath19wayne state university , detroit , michigan 48202 + @xmath20california institute of technology , pasadena , california 91125 + @xmath21university of california , san diego , la jolla , california 92093 + @xmath22university of california , santa barbara , california 93106 + @xmath23carnegie mellon university , pittsburgh , pennsylvania 15213 + @xmath24cornell university , ithaca , new york 14853 + @xmath25university of florida , gainesville , florida 32611 + @xmath26harvard university , cambridge , massachusetts 02138 + @xmath27university of illinois , urbana - champaign , illinois 61801 + @xmath28carleton university , ottawa , ontario , canada k1s 5b6 + and the institute of particle physics , canada + @xmath29ithaca college , ithaca , new york 14850 + @xmath30university of kansas , lawrence , kansas 66045 + @xmath31university of minnesota , minneapolis , minnesota 55455 + @xmath32state university of new york at albany , albany , new york 12222
globular clusters ( gcs ) are the oldest and most massive stellar aggregates in our galaxy . as such , they are the best natural laboratories to study large , co - eval populations of stars at known distance and metallicity . indeed , much of our understanding of star formation and evolution has been derived from observational studies of gcs . nonetheless , we still lack an understanding of the very low mass stars ( vlmss ) around the faint end of the hydrogen - burning main sequence ( ms ) and of objects beyond that limit , i.e. , sub - stellar sources , so called `` brown dwarfs '' ( bds ) . this is especially true for objects at the low - metallicities typical for the gcs in our milky way ( see below ) . bds present a link between stars and planets , and thus are important for our understanding of both star and planet formation and evolution . bds are sub - stellar objects that are not massive enough to ignite and sustain hydrogen burning . thus , where low - mass stars will retain their luminosity for a hubble time or longer , a bd will continue to cool and become fainter with age . like giant gas planets , bds have complex atmospheres @xcite . the distinction between stars , bds and planets is either based on mass or on formation . in general , stars have masses @xmath4 and can sustain hydrogen burning , bds have masses between 80 and 13 @xmath5 and can not sustain hydrogen burning but a short period of deuterium burning , and giant planets have masses below 13 @xmath5 and can not sustain deuterium burning ( burrows et al . 1997 , stamatellos 2014 , but see also sect . [ hmass ] ) . because low - mass stars and bds have life - times much longer than the age of the galaxy , the gc vlmss and bds are also important tracers of galactic formation and chemical evolution . the formation of bds is a matter of considerable dispute . they might have formed in the same way as ( low - mass ) stars from turbulent cloud fragmentation @xcite , which would imply a continuous extension of the imf into the sub - stellar regime . on the other hand , bds might form e.g. from the ejection of stellar embryos or sub - stellar clumps which did not have the chance to accumulate enough mass @xcite . kroupa & bouvier ( 2003 ) suggested that bds form via photo - evaporation of protostars through nearby massive stars . this might also suggest an increase in the number of bds with cluster mass , as more massive clusters have more o stars which can produce more bds . bds might also form from the fragmentation of circumstellar disks ( stamatellos et al . 2011 , thies et al . 2010 , kaplan et al . 2012 ) . in this case , the number of bds in clusters could be enhanced in dense clusters as dynamical interactions between cluster stars lead to more disk fragmentation . thies et al . ( 2015 ) concluded that bds likely form not just via one formation scenario , but from a combination of various channels . large surveys undertaken in the past decade have detected large numbers of bds . for example , the two micron all sky survey ( 2mass ; skrutskie et al . 2006 ) , the sloan digital sky survey ( sdss ; york et al . 2000 ) , the united kingdom infrared telescope deep sky survey ( ukidss ; lawrence et al . 2007 ) , the wide - field infrared survey explorer ( wise , wright et al . 2010 ) all sky survey have been very successful in finding such cool , low - mass objects . however , for most of these bds , key physical properties like metallicity and age are unconstrained ( see e.g. ) . in fact , determining the physical parameters is extremely difficult and a major hurdle in bd research . observations of open star clusters and star forming regions , where all sources are at the same distance and metallicity , can mitigate this problem ( e.g. steele et al . 1995 , rebolo et al . 1996 , martn et al.2001 , pinfield et al . 2003 , boudreault & lodieu 2013 , casewell et al . 2014 ) , but only for young and metal - rich objects . thus the need for benchmark sources is especially evident for the metal - poor regime , and indeed we still do not know much about _ old , metal - poor _ bds . so far , only very few low - metallicity , old vlmss near the h - burning limit , and even fewer sub - stellar ( candidate ) halo objects have been identified ( e.g. burgasser et al . 2003 , 2009 , lpine et al . 2004 , burgasser & kirkpatrick 2006 , cushing et al . 2009 , sivarani et al . 2009 , murray et al . 2011 , mace et al . 2013 , pinfield et al . 2014 , luhman & sheppard 2014 , burningham et al.2014 , kirkpatrick et al . 2014 ) . this is where gcs come in : they are massive , and thus might have produced bds in large numbers , and they are also the oldest and most metal - poor stellar aggregates in our galaxy . potentially , gcs are the ideal hunting ground for old , metal - poor benchmark vlmss and bds which are much needed if we are to test stellar and sub - stellar formation and evolution theories and models of metal - poor ( sub-)stellar atmospheres . however , identifying substellar objects in gc is challenging due to their intrinsic faintness . therefore , the closest gcs make the best targets for this kind of research . out of the gcs in our galaxy , m4 ( ngc6121 ) is the closest gc to us ; distance estimates range from 1.7 kpc ( hansen et al . 2004 ) to @xmath6 2 kpc ( e.g. bedin et al.2009 ) . braga et al . ( 2015 ) estimated a true distance modulus of 11.28 mag based on rr lyrae period - luminosity and period - wesenheit relations , resulting in a distance of 1.8 kpc , which agrees well with previous estimates . malavolta et al . ( 2014 ) analyzed 7250 spectra for 2771 cluster stars and found a metallicity of @xmath7=-1.07 $ ] dex ( rgb ) to @xmath7=-1.16 $ ] dex ( sub giant branch and ms stars ) , which is well in line with previous metallicity estimates ( e.g. @xmath7=-1.15 $ ] dex according to the 2010 update of the harris , 1996 , catalogue of globular clusters in the milky way ) . as such , m4 is a prime target for ultra - deep observational studies . indeed , deep optical studies with the hubble space telescope ( _ hst _ ) have been undertaken by richer et al . ( 1997 , 2004 ) and bedin et al . ( 2009 ) , yielding impressive results . richer et al . ( 1997 , 2004 ) estimated the fraction of similar - mass photometric binaries to be small ( just 2% in their outer field , falling to just 1% towards the cluster core ) . milone et al . ( 2012 ) suggested a much higher total binary fraction , raising from 10% at the halfmass radius to 15% towards the cluster core . the present day mass function of the lower - mass ms stars is flat ( bedin et al . 2001 ) , with a slope of @xmath8 and a further flattening towards the cluster centre ( richer et al . the cluster wds suggest that the initial mass function ( imf ) above 0.8 @xmath9 was much steeper than the present day mass function . bedin et al . ( 2009 ) presented the deepest optical colour - magnitude diagram ( cmd ) to date of this cluster and located the faint end of the wd cooling sequence in m4 at @xmath10 mag , suggesting an age of [email protected] gyr . this agrees with the finding of hansen et al . ( 2004 ) , who found a wd based age of 12.1 gyr . the ongoing _ hst _ m4 core project ( pi l. bedin , go-12911 ) searches for binary dark companions to ms stars . the high - accuracy astrometry and photometry of this data - set , together with archival material and a part of the deep near - ir ( nir ) data presented in this paper , have already been used to identify two distinct sequences along the lower - mass ms @xcite . m4 is also so far the only gc known to host a planetary system , psr b1620 - 26 ( sigurdsson et al . 2003 ) , which challenged the planet - metallicity relation of the standard planet formation model at that time @xcite . recently , hasegawa & hirashita ( 2014 ) suggested that the critical metallicity for gas giant formation is [ fe / h]@xmath12 dex , which agrees with m4 s metallicity . beer et al . ( 2004 ) suggested a metallicity - independent formation scenario , in which the planet in m4 formed through dynamically induced instability in a circumbinary disc . if true , then we can expect many planets to form especially in the dense gcs in which dynamical interactions between cluster stars are ubiquitous . since bds might form in a similar way , we then might also expect that many more bds form in dense gcs compared to open clusters . this paper is structured as follows . in section [ obs ] we describe the observations and reduction of our nir and archival optical data . in section [ cmds ] we present and discuss the nir and optical - nir cmds , and our results and conclusions are summarized in section [ summary ] . the nir observations of the globular cluster m4 were carried out in april 2012 with the wide field camera 3 ( wfc3 ) on board the _ hst _ , using the @xmath3 and @xmath13 filters ( program go12602 , pi : dieball ) . all observations were made at a single pointing position on a field centred at about one core radius north east from the cluster centre . this region in the cluster had been the focus of the programs go-5461 ( richer et al . 1997 , 2004 ) and go10146 ( bedin et al . 2009 ) and thus is fully covered by deep optical observations . a standard 4-point wfc3-irdither - box - min dither pattern and a sampling of nsamp14 spars50 was applied during the observations to get a well - sampled point spread function ( psf ) . the wfc3 field of view is @xmath14 in the ir channel , with a resolution of @xmath15 per pixel . all nir observations were carried out during two consecutive orbits on april 16th ( @xmath3 ) and four orbits on april 20th to 21st ( @xmath13 ) , comprising a total of 8 individual images in @xmath3 and 16 images in @xmath13 , each 653 seconds , resulting in total exposure times of 5223 seconds ( @xmath3 ) and 10447 seconds ( @xmath13 ) . using the pipeline produced flat - fielded ( flt ) images , we first created a master image for each of our ir filters , using multidrizzle running under pyraf ( the python - based interface to iraf ) . multidrizzle corrects geometric distortions that are present in the input images and combines them into a master image . shifts between the individual images are expected , as we have applied a dither pattern . on top of this , telescope breathing can affect guide star tracking and as a result can cause small shifts , typically on a sub - pixel scale ( see e.g. the multidrizzle handbook available on the stsci webpages ) . in order to ascertain that all shifts are taken into account , we created geometric - distortion corrected individual images . based on the coordinates of the same 10 stars in each of these images ( selected to be well distributed over the field of view ) , accurate shifts between the individual and the master images were then determined using tweakshift . the master images created in this way are displayed in figures [ f110 ] and [ f160 ] . note that these master images serve as reference images for the positions of the stars detected by dolphot , but the photometry was actually performed on the individual images ( see [ dolphot ] below ) . photometry was performed on the individual flt images using the dolphot software developed by a. dolphin as a generalized version of hstphot @xcite . dolphot runs on the flt images downloaded from the stsci archive , i.e. , no further processing of the images is necessary nor recommended . photometry is performed on each of the input images , using the wfc3 module which replaces the analytic psf model with a look - up table computed using tiny tim psfs ( krist & hook 2011 ) . for our data , we used jay anderson s psf libraries for the wfc3/ir filters and the pixel area maps available from the dolphot webpage . the wfc3 module also includes a photometric calibration to the vegamag system . as a reference frame for a common physical ( image ) coordinate system we use our deepest image , i.e. , the master image created with multidrizzle . the dolphot package also provides routines to mask all pixels that are flagged as bad in the data quality arrays , to multiply with the pixel area maps , to calculate sky images , and to align all input images ( or rather the source coordinates found in each input image ) to the reference frame ( our drizzled master images ) using user defined lists of stellar coordinates in each input and the reference image . for these steps , we used the recommended wfc3 ir parameter settings . note that performing photometry on the drizzled image is expected to provide sub - optimal photometry , because the drizzling process affects the psf and the noise characteristics of the drizzled images . instead , dolphot runs on all input images simultaneously and thus is capable of providing deep photometry . we started with the recommended parameters , and then refined some parameters to push our photometry as deep as possible . dolphot can run on data from multiple filters and cameras , thus we were able to do the photometry on both @xmath3 and @xmath13 data sets simultaneously . the output file includes x and y positions , photometric parameters like sharpness , crowding ( the magnitude difference to the measured magnitude if no neighbouring sources would be fit simultaneously ) , the object type , and magnitudes in the vegamag system for all sources detected in the @xmath3 and @xmath13 data . a calibrated nir cmd is plotted in fig . [ cmdirall ] . as can be seen , the cmd is exceptionally deep . the total catalogue contained 51311 sources , but a large number of those will be spurious detections . + among the archival _ hst _ material the only images which offer any hope of detecting bds in the optical are those deep and in the reddest filters available , i.e. , @xmath16 and @xmath17 ( indeed , technically @xmath182/3 of these pass - bands are already in the nir wavelength range ) . we extensively searched in the _ hst _ archive , and identified four @xmath191200 seconds deep images taken with acs / wfc in @xmath16 under program go-10146 ( pi : bedin ) as the optimal for our purposes , as those images were taken in low - sky mode , are well dithered , and collected in a well defined epoch in 2005.48 ( i.e. , about 6.8 years before the go-12602 data ) . images were reduced with the software described in great detail in anderson et al . briefly , the method is essentially a psf - fitting where all pixels from all images are simultaneously fitted using the appropriate psfs which account for the spatial and temporal shape - variation on each individual image . the key to the method is an optimal knowledge on how to transform those pixels into a common reference frame and an exquisite empirical psf modeling . this software is well tested and used in all the twelve works of the series `` the acs survey of galactic globular clusters '' ( see , sarajedini et al.2007 ) . the first photometric run returned 19990 optical detections . as the detection of bds might be only marginal in these optical images , we relaxed the finding criteria in a second photometric run , imposing to save all the local maxima , and not only the significant ones , resulting in 1.5 million detections . although most of these local maxima likely are just fluctuations of the background noise , we can still use them to set an optical upper limit to the nir - detected bd candidates . for visual inspection of the wds and bd candidates , stacked images were created from the cte - corrected flt images . the stacked images have pixels values resulting from median clipping of the cte - corrected flt individual images , and are supersampled by a factor of 2 in each direction ( see anderson et al . 2008 and anderson & bedin 2010 for further details ) . we used the data in the two nir filters @xmath3 and @xmath13 to create a nir cmd . as the full nir catalogue contains a large number of spurious detections , we created a best - photometry nir cmd ( described below in sect . [ nircmds ] ) . however , since wds and bd candidates have similar magnitudes and colours in the nir , bd candidates can not be distinguished based on the nir data alone . we then used additional optical f775w data to create proper motion cleaned optical - nir cmds ( see sect . [ nircmds ] ) in which the wd and ms sequence are clearly separate . thus , the optical data are used to distinguish the wds from the bd candidates . in order to produce a clean cmd of only `` good '' stellar sources , we selected the output catalogue in such a way that the faint sources will be dominated by true stellar detections while keeping the number of spurious detections at bay . * * the resulting best - photometry catalogue includes 2526 sources , i.e , @xmath20 of all detections . the best - photometry cmd is remarkably clean , as can be seen in fig . [ cmdirsel ] . we can clearly see the ms delineating down towards the expected h - burning limit and going beyond that to fainter sources . note that this cmd is not proper motion cleaned , as we show all nir sources that satisfy our selection criterion on the photometry . all sources with magnitudes fainter than 24 mag in @xmath3 were visually inspected on the @xmath3 master image , resulting in 177 visually confirmed nir sources around or fainter than the h - burning limit in @xmath3 ( see sect [ hmass ] below ) . the ms is narrow between the two ms `` knees '' ( @xmath21 mag and @xmath22 mag ) . the first `` knee '' occurs around @xmath23 k and a mass of @xmath24 and is due to the formation of molecules in the cool stellar atmospheres , and the second knee is a result of increasing electron degeneracy in the ( sub)-stellar interior close to the h - burning limit ( baraffe et al . below the second knee , the ms broadens towards fainter and lower - mass ms stars . indeed , the low - mass ms splits into two branches , as shown in milone et al . ( 2014 ) , who had first pointed out multiple stellar generations among vlmss in m4 , based on high - precision deep optical _ data from the _ hst _ m4 core project and our nir data . the split can be clearly seen in the nir cmd ( see milone et al . 2014 , their fig . 2 ) , as opposed to the optical data , demonstrating that the nir is also an ideal waveband to search for multiple sequences along the lowest - mass and hence faintest ms . note that the goal in milone et al . ( 2014 ) was to search for multiple generations along the low mass ms based on very precise photometry . in contrast , in this paper , our goal is to go deep and well beyond the h - burning limit . in our fig . [ cmdirsel ] , we can see the ms delineating further down towards the expected end of the h - burning sequence , marked with red slashed lines and light - red shaded area . see sect . [ hmass ] below for a discussion on the mass and nir magnitude at the h - burning limit . around @xmath25 mag the number of sources on the ms decreases and the ms peters out . on the blue side of the faint ms , the wd sequence can be seen , starting around @xmath26 mag and @xmath27 mag and going fainter and redder . an increase in source number seems to be apparent below @xmath28 mag , i.e. , below the expected end of the h - burning sequence . this is the area in the cmd where we expect the bds to appear . this happens to coincide with the wd sequence as well , i.e. , this is the region where wd and bd cooling sequences would be expected to cross . unfortunately , this also means that we can not disentangle wds and bd candidates based on our nir data alone . in order to help with both cluster membership determination and distinguishing wds and bd candidates , the deep optical data from go10146 were used ( see sect . [ optobs ] and sect , [ optnircmds ] ) . the wd as well as the bd regions have been indicated in fig . [ cmdirsel ] . for orientation purposes , we have plotted a wd sequence ( blue line ) . the wd cooling sequence was constructed by interpolating on the wood ( 1995 ) grid of theoretical wd cooling curves , adopting a mean wd mass of @xmath29 . using a grid of synthetic da wd spectra kindly provided by b. gnsicke ( see gnsicke et al . 1995 ) we carried out synthetic photometry with pysynphot . note that we have shifted the wd cooling sequence to get a reasonable match to the underlying cmd . this required a rather large distance of 2.2 kpc and a reddening of @xmath30 mag ( a standard reddening law ( seaton 1979 ) is built into pysynphot ) . our cooling sequence starts at @xmath31 k and terminates at @xmath32 , but note that the coolest wds in m4 have temperatures as low as @xmath33 ( bedin et al . 2009 ) . in addition , we have marked the location of the known field bd sdss - j125637.13 - 022452.4 @xcite , which has a metallicity similar to m4 but is likely several gyr younger . as a consequence , its cooling time is shorter and thus it is expected to be brighter than the m4 bds . the observed @xmath34- and @xmath35-band magnitudes agree with a 5 gyr old , @xmath36 source at a metallicity of [ m / h]=-0.5 dex . we scaled the observed @xmath34- and @xmath35-band magnitudes to the wfc3 nir filters and applied m4 s distance and reddening ( following hendricks et al . 2012 , we adopted a reddening of @xmath37 mag , a true distance modulus of 11.28 mag , @xmath38 , @xmath39 and @xmath40 ) and over - plotted the field bd on our cmd . its location supports that our data are indeed deep enough to reach well into the bd zone . two 12 gyr isochrones have been overplotted on the best - photometry cmd , a bt - settl model based on the asplund et al . ( 2009 ) solar abundances and the barber & tennyson ( 2006 ) line list ( allard et al . 2012 ) , for a metallicity of @xmath41}=-1 $ ] dex ; and a dartmouth model for @xmath42}=-1.2 $ ] and @xmath43}=+0.4 $ ] ( dotter et al . note that we do not attempt to derive cluster parameters from the isochrone fitting , instead , we have fit the isochrones by eye so that they best overlap with the underlying cmd for the dartmouth isochrone . for the bt - settl isochrone , we chose a smaller distance modulus of 11 mag and @xmath44 . the difference in shape of the isochrones , as well as the difference in the best - fit parameters , reflect the differences in the underlying physics , i.e. treatment of the stellar atmospheres including molecules . for a more in - depth discussion of the input physics to the models we refer the reader to the bt - settl and dartmouth webpages and the references given there ( https://phoenix.ens-lyon.fr/grids/bt-settl/ and http://stellar.dartmouth.edu/models/ ) . ] . note also that both sets terminate at a stellar mass of 0.083 @xmath9 ( bt - settl ) or 0.1 @xmath9 ( dartmouth ) , i.e. , they do not reach to the h - burning limit . our nir and optical catalogues have been matched using a six - parameter linear transformations between the star positions in the different epochs . as reference stars for the transformations we used only well - measured , isolated , non - saturated cluster stars with a high signal - to - noise and low residuals . the _ predicted positions _ of the first epoch sources in the second epoch are compared with the _ observed positions _ and the displacements between first and second epoch are calculated . the top panel in figures [ cmdj ] , [ cmdh ] , [ pmcmdir110 ] and [ pmcmdir160 ] shows the displacements in wfc3/ir , based on the total dolphot nir ( 51311 detections ) and optical ( 19990 sources ) catalogue . since cluster stars have been used for the reference list , we expect cluster members to agglomerate around zero in the displacement vector point diagram . indeed , two populations can be distinguished : a dense and tight agglomeration of data points around @xmath45 and @xmath46 which denotes the cluster members , and a more widely spread data region centering around @xmath47 and @xmath48 which denotes field sources . the latter are mostly bulge sources , reflecting the low tangential motion of m4 around the galactic center ( see also bedin et al . 2003 ) . the corresponding cmds are plotted in the second row in figures [ cmdj ] , [ cmdh ] , [ pmcmdir110 ] and [ pmcmdir160 ] . the left cmds show all sources within up to two wfc3 pixels displacement , the middle cmds show only sources with a displacement of no more than 0.1 pixels which suggests that they are cluster members , and all non - cluster sources are shown in the right diagrams . a displacement of 0.1 pixels corresponds to a proper motion of 1.9 mas / year , based on our timeline of 6.8 years and a pixel scale of 0.13@xmath49 . this agrees well with e.g. bedin et al . ( 2003 , 2009 ) and zloczewski et al . as can be seen , the proper - motion cleaned `` cluster '' cmds ( middle diagrams ) show a well defined ms , terminating around @xmath50 mag in both cmds in figures [ cmdj ] and [ cmdh ] , as well as a well defined wd sequence ( light - blue data points ) going down to the bottom of the wd sequence around @xmath2 mag . on the other hand , the proper - motion cleaned optical - nir cmds do not show any bd candidates , i.e. , sources fainter than @xmath51 mag and redder than @xmath52 or @xmath53 mag ( see also fig . [ models ] ) this was expected , as the deep optical cmd presented in bedin et al . ( 2009 ) did not show any potential ms sources fainter than the h - burning limit . however , and most importantly in the context of this paper , the bds are expected to be much redder than the wds in the optical - nir cmds . indeed , the wd sequence and the ms are clearly separated in our optical - nir cmds . therefore we can use the deep optical data set to identify the wds in our nir cmds and disentangle wds from potential bds . what is the minimum mass at the hydrogen - burning limit in a low - metallicity cluster like m4 ? and , as a consequence , at which nir magnitude and colour do we expect the h - burning limit in our nir cmd ? early theoretical work ( kumar 1963 ) suggested a lower limit to the stellar ms at @xmath54 for population i and @xmath55 for the metal - poor population ii stars , to which also gcs belong . hayashi & nakano ( 1963 ) suggested that stars less massive than @xmath56 can not undergo hydrogen burning . they further suggested that the limiting mass is not very different for population i and population ii stars . burrows et al . ( 1993 ) presented a zero metallicity theoretical model ( i.e. , population iii ) which suggests a limiting h - burning mass as high as @xmath57 . treatment of the atmosphere has a considerable impact on the predicted limiting mass , as can be seen in chabrier et al . ( 2000 ) who suggested a limiting mass of @xmath58 for models that include dust formation , saumon & marley ( 2008 ) who suggested a limiting mass of @xmath59 for cloudless models , and @xmath60 for cloudy models ; all for solar metallicity and an age of 10 gyr . previous theoretical works suggest that the h - burning limit is at higher masses for more metal - poor stars . the models roughly agree on a limiting mass of @xmath59 for solar metallicities . following hayashi & nakano ( 1963 ) , we conservatively assume that the h - burning mass for a population as metal - poor as m4 is between @xmath59 and @xmath56 . unfortunately , detailed sub - stellar models for sub - solar metallicities are presently not available . updated models that extend well into the bd regime are currently being computed ( allard , private communication ) . however , the most recent set of bt - settl models ( allard et al . 2012 ; 2013 ) suggest a strong metallicity dependence of the shape and luminosity of the low - mass ms . these models only go down to @xmath61 and are close to the h - burning limit ( allard private communication ) , but do not go beyond the stellar sequence into the sub - stellar regime . thus , we linearly extrapolated the bt - settl models ( based on the asplund et al . 2009 solar abundances ) for @xmath62=-1 $ ] dex and an age of 12 gyr ( closest to m4 s parameters ) down to sub - stellar masses of @xmath63 , and applied distance and reddening as in fig . [ cmdirsel ] for the bt - settl model . the extrapolated models are plotted in fig . [ models ] , and the magnitudes around the h - burning limit are listed in table [ mlim ] . as mentioned above , no sub - stellar models for low metallicities are currently available . different models exist for a metallicity of @xmath41 } = 0.0 $ ] dex . thus , in fig . [ compare ] we compare our extrapolated models with these more metal - rich models , all for an age of 12 gyr and scaled to m4 s distance and reddening . the effect of the different atmospheric physics can be clearly seen in the shape and colour of the models . however , the expected end of the h - burning sequence between a @xmath59 and a @xmath56 is at comparable magnitudes in all sub - stellar models . this gives us some confidence that we can use the extrapolated metal - poor bt - settl models to get an estimate of the magnitude and colour range of the h - burning limit . for more exact values we will have to wait for the updated metal - poor models , but we remind the reader that the main purpose of this project is to provide the metal - poor benchmark sources and thereby fill the observational plane with data points that are needed to constrain theoretical models . unlike stars , bds can not retain their luminosities via nuclear fusion . as a consequence , they cool with time , and a bd will be at fainter luminosities in an old gc compared to a young bd of the same mass and metallicity . to get an estimate of this effect , we used nextgen models ( baraffe et al . 1997 , baraffe et al . 1998 ; see fig . [ bdcool ] ) . in the left diagram , we plot mass against @xmath64 for various ages . the right diagram shows 1 gyr and 10 gyr nextgen models for different metallicities . since substellar isochrones at m4 s low metallicity of [ m / h]=-1 dex and ages of @xmath65 gyr currently do not exist , we also plot isochrones for a metallicity of [ m / h]=-0.5 and 0.0 dex which go down to lower masses . as can be seen , metallicity has a considerable impact on the shape of the isochrones as well as on the cooling time scale and hence fading of low - mass , substellar objects . the [ m / h]=0 dex isochrones continue to extend with time to fainter magnitudes and redder colours , suggesting that a 0.075 @xmath66 source fades by @xmath67 mag in @xmath3 from 1 to 10 gyr , and a 0.08 @xmath66 source becomes fainter by @xmath68 mag . the [ m / h]=-0.5 dex isochrones suggest that a source of 0.079 @xmath66 becomes fainter by @xmath67 mag , and a 0.08 @xmath66 source becomes fainter by @xmath69 mag , but the sources also become bluer rather than redder . the [ m / h]=-1 dex 10 gyr isochrones terminates at 0.083 @xmath66 . according to the models , such a metal - poor , low - mass source already fades by 0.3 mag in @xmath3 from 1 to 10 gyr . the models suggest that the blue - turn is more pronounced for lower - metallicities , and also metal - poor sources become fainter compared to metal - richer sources at the same mass , i.e. they cool faster . our optical - nir cmds presented in figures [ cmdj ] and [ cmdh ] show that the observed ms peters out around @xmath50 mag , suggesting that we are approaching the end of the h - burning sequence . our nir cmd in fig . [ cmdirsel ] suggests that the ms peters out just below @xmath70 mag where the density of stars decreases . at fainter magnitudes , the wd sequence crosses the ms and the star density increases again . based on the cmds , we estimate our detection limits around @xmath2 mag , @xmath0 mag and @xmath1 mag . the detection limits are also indicated in fig . [ models ] with a dotted line . comparing the theoretical models in fig . [ models ] with our observed cmds and taking the detection limits and the predicted h - burning limit into account , we conclude that we have reached beyond the h - burning limit in our nir cmd and are probably just above or around this limit in our optical - nir cmds . cccc mass [ @xmath9 ] & @xmath3 [ mag ] & @xmath13 [ mag ] & @xmath16 [ mag ] + 0.075 & 25.170 & 24.736 & 30.105 + 0.077 & 24.771 & 24.281 & 29.327 + 0.080 & 24.172 & 23.598 & 28.160 + as with the best - photometry faint nir sources , we visually inspected all cluster wd candidates , i.e. , sources whose proper motion ( or rather displacements ) suggest that they are cluster members and whose position in the optical - nir cmds suggest that they are wds . our first photometric run on the optical data did not return an optical counterpart for 59 of the best - photometry faint nir sources . we then over - plotted the positions of those 59 faint nir sources ( i.e. , _ without _ an initial optical counterpart ) on the optical @xmath16 master image and inspected each position by eye . visual inspection of the @xmath16 master image showed a faint optical source at or close to the location of the nir source in most cases . thus , for the second optical photometric run ( see sect . [ optobs ] ) , the parameter settings were relaxed so that all local maxima were retained , resulting in 1.5 million detections . nearly all of those are just spurious detections ( i.e. , background noise or spikes in psf wings that are not real stellar sources ) , however , a further 47 optical counterparts to the faint best - photometry nir sources were detected and thus were added to the initial list of optical - nir matches . for the remaining twelve faint nir sources , still no optical counterpart was returned . however , out of those , five are located on psf streaks and two in the acs wfc chip gap , so that nothing can be said about an optical counterpart . for four of the remaining five sources , visual inspection of the @xmath16 master image seem to indicate an optical source on the position of the nir source , but no photometric measurement was possible . however , we used nearby stars of similar brightness ( based on pixel counts ) to estimate magnitude limits . one of these sources appears to have an optical counterpart in the centre of our search circle , probably at @xmath50 mag . this is too bright for a bd candidate and thus makes this source a wd candidate . thus , we do not consider it further . the remaining four sources , i d 1 to 4 , with no optical photometry are listed in table [ bdcand ] , and images are shown in fig . [ bdimas ] . for comparison , we also show images of wds selected from the proper motion cleaned cmd and which have similar nir magnitudes to the four nir sources without optical photometry ( see fig . [ wdimas ] and table [ wdcand ] ) . at the rim of our search circle for source 1 , an optical source can be seen . this , however , is too far away ( @xmath72 away from the position of the nir source ) and would not agree with being a cluster member . in the centre of our search circle , an optical source just might be visible . if true , this source would have @xmath73 mag , i.e. , the local detection limit . for source 2 , an optical counterpart is visible in the centre of the search circle , again this source would be close to the detection limit at @xmath73 mag . thus , both source 1 and 2 are likely ( massive ) bd candidates . unfortunately , source 2 is close to a saturation streak from a nearby bright stars in the optical image . the potential counterpart , however , can be clearly distinguished . source 3 shows a `` bright '' optical source at the rim of our search circle , probably at @xmath74 mag , again based on the magnitudes of nearby stars that appear to be of similar brightness . however , this optical source is @xmath75 away from the position of the nir source and thus too far away to agree with being a cluster member . a very faint optical source might just be in the centre of the search circle , if true , then this optical source would be at @xmath76 mag , the local detection limit , making source 3 a good bd candidate . just one source , source 4 , does not show an optical counterpart at the location of the nir source ( i.e. , within our search circle ) , and is thus our best bd candidate . since the optical photometry did not return magnitudes fainter than @xmath77 mag , an optical counterpart to bd candidate 4 must be fainter than this absolute optical limit . note that source 4 appears somewhat extended and could probably consist of two or three faint sources , or possibly an extended object ( although we expect galaxies to be much bluer , see bedin et al . 2009 ) . the positions of the four bd candidates are indicated in the cmds in fig . [ pmcmds ] . since we do not have an optical measurement , we provide the upper optical magnitude limits given in table [ bdcand ] . we also overplot the extrapolated bt - settl 12 gyr isochrone , using the distance and reddening parameters derived from the best fit in the nir cmd . as can be seen , all four bd candidates are very close to the extension of the ms into the bd regime , which supports our classification of these sources as good bd candidates . source 2 is blueward of the isochrone . its position in the optical - nir cmds agrees with this source being a faint wd at the very bottom of the ( optical ) wd cooling sequence in m4 , but it also agrees with this source being a vlms star or a massive bd . in the nir cmd this source is _ not _ at the very bottom of the wd sequence . we suggest that this source is a good bd candidate . sources 1 , 3 and 4 all are very close to the ms , and their position in the optical - nir cmds does not suggest that these sources are wds , but rather bds . wds are much bluer and brighter than our bd candidates . lllllll id@xmath78 & ra & dec & @xmath3 & @xmath13 & @xmath16 & comment + & [ h : m : s ] & [ @xmath79 & [ mag ] & [ mag ] & [ mag ] & + 1 & 16:23:41.701 & -26:29:17.82 & @xmath80 & @xmath81 & @xmath8228 & bd candidate + 2 & 16:23:45.443 & -26:30:05.58 & @xmath83 & @xmath84 & @xmath8228 & wd / bd candidate + 3 & 16:23:44.711 & -26:29:30.90 & @xmath85 & @xmath86 & @xmath87 & bd candidate + 4 & 16:23:45.371 & -26:29:37.40 & @xmath88 & @xmath89 & @xmath1832.9 & bd candidate + llllll id@xmath90 & ra & dec & @xmath3 & @xmath13 & @xmath16 + & [ h : m : s ] & [ @xmath79 & [ mag ] & [ mag ] & [ mag ] + 1 & 16:23:44.354 & -26:30:26.09 & @xmath91 & @xmath92 & @xmath93 + 2 & 16:23:36.848 & -26:29:33.58 & @xmath94 & @xmath95 & @xmath96 + 3 & 16:23:36.953 & -26:29:33.81 & @xmath97 & @xmath98 & @xmath99 + 4 & 16:23:47.062 & -26:30:37.80 & @xmath100 & @xmath101 & @xmath102 + figures [ pmcmdir110 ] and [ pmcmdir160 ] again show the nir cmds , but only for those nir sources for which an optical counterpart had been found ( black data points ) . the middle cmds show only nir sources whose optical counterparts agree with being cluster members , based on the displacement vector point diagram . we also show best - photometry nir sources _ without _ optical counterparts , plotted in red . as mentioned in sect . [ nircmds ] , all 177 best - photometry nir sources fainter than @xmath103 mag have been inspected on the @xmath3 and the optical @xmath16 master images . out of these faint 177 nir sources , 165 have an optical counterpart . out of these 165 faint nir / optical sources , 48 are cluster members based on their proper motion ( displacements ) , and are located on the wd sequence . the remaining five faint best - photometry nir sources which are not on psf streaks or on the chip gap , and _ without an optical counterpart _ ( and hence without a proper motion estimate ) , all make good bd candidates to start with . visual inspection of these five sources and estimates on the optical magnitude limit ( see above ) suggests that one is probably a wd candidate , and the remaining four sources , listed in table [ bdcand ] , are good bd candidates . how many bds can we expect ? this number is highly uncertain and depends on the assumed bd formation scenario ( see sect . [ intro ] ) . furthermore , given our detection limits we can only expect to find the most massive bds with masses larger than 0.068@xmath9 ( based on the extrapolated bt - settl models ) . richer et al . ( 2004 ) derived a rather flat present - day mass function for m4 . extrapolating towards fainter and lower - mass stars , they estimate that between 15 to 50 vlmss with masses between 0.085 and 0.095 @xmath9 should be in their field of view ( go-8679 , wfpc2 data ) . using the theoretical models ( fig . [ models ] ) we can count the number of vlmss in our field . if we only consider nir sources that have an optical counterpart and a proper motion that suggests that they are cluster members , and best - photometry nir sources without an optical counterpart , we find 23 vlmss in a mass interval between 0.08 and 0.09 @xmath9 . assuming that the mass function is flat and that the slope of the mass function does not change considerably across the stellar / sub - stellar border , we can expect a similar number of bds with masses between 0.070 and 0.08 @xmath9 . however , the number of bds formed per star is probably more around @xmath104 ( see e.g. thies et al . 2007 ) . in this case , we can expect @xmath6 5 bds . this is of course a very rough estimate , but it does agree with our finding of four bd candidates . how many foreground or background sources can we expect in our cluster cmd ? we can simply count the number of sources that are well outside the area covered by the cluster and the clump of field stars around @xmath47 pixels and @xmath48 pixels in the vector point diagram . we find a field star density of 165 field stars per @xmath105 . scaling this number to the area covered by our cluster stars , i.e. , a displacement of 0.1 wfc3 pixels , we find that we can expect 5.2 field stars in our cluster cmds . how many foreground stars can we then expect to have nir magnitudes and colours similar to our bd candidates ? selecting only field stars with @xmath106 mag and @xmath107 mag , the field star density is reduced to 50 stars/@xmath105 , and scaled to the area covered by the cluster in the vector point diagram , we then find that 1.6 such sources can be expected among our pm selected faint cluster members . thus , about half of our suggested bd candidates might actually be foreground or background sources that happen to move with the cluster velocity across the plane of the sky . we have presented the deepest nir _ hst_/wfc3 study of the gc m4 to date . the nir data were proper - motion cleaned using archival deep optical _ hst_/acs ( @xmath16 ) data . our best - photometry nir cmd reveals a narrow ms delineating down towards the expected end of the h - burning sequence . 177 best - photometry nir sources fainter than the h - burning limit in @xmath3 ( @xmath108 mag ) could be identified in our @xmath3 master image . for 165 of these faint nir sources , an optical counterpart was found , 48 of these are cluster members according to their proper motion . all of these 48 faint cluster sources are on the wd sequence . we found in total five faint nir sources for which the optical photometry did not return a measurement ( and which are not on psf streaks or on the chip gap ) . we then visually inspected the positions of these faint nir sources on the optical images and estimated , where possible , upper optical magnitude limits of potential optical counterparts that just might be visible . one source is likely another wd and rejected as a bd candidate . based on the upper optical magnitude limits , we indicate the position of the remaining four sources in the optical - nir cmds . one of the sources ( source 4 in table [ bdcand ] ) , does not show an optical counterpart at all , which implies that its optical counterpart must be fainter than the absolute optical detection limit of @xmath109 mag . this source appears to be somewhat extended in the nir image , which might indicate multiple faint sources , i.e. multiple bds , or possibly a galaxy . however , its position in the cmds does agree with this source being a bd . one source ( source 2 in table [ bdcand ] ) might be another wd candidate , but its position in the optical - nir cmd also agrees with this source being a massive bd or a vlms star at the bottom of the ms . the remaining two sources also have positions that indicate that these sources are massive bds . we conclude that we have found four good bd candidates , but we caution that further studies and deeper optical data are necessary to confirm their status and cluster membership . we are grateful to an anonymous referee for her / his valuable comments which helped to improve this paper . thanks andrew dolphin for helpful discussions about dolphot . this work was supported by nasa through grant go-12602 from the space telescope science institute , which is operated by aura , inc . , under nasa contract nas5 - 26555 . a.d . acknowledges support from the people program ( marie curie actions ) of the european union s seventh framework program fp7-people-2013-ief under rea grant agreement number 629579 . l.r.b.acknowledges prin - inaf 2012 funding under the project entitled : `` the m4 core project with hubble space telescope '' . allard f. , homeier d. & freytag b. 2012 , rspta , 370 , 2765 allard f. , homeier d. & freytag b. 2013 , mmsai , 84 , 1053 anderson j. , sarajedini a. , bedin l. r. , king i. r. , piotto g. et al . 2008 , aj , 135 , 2055 anderson j. & bedin l. r. 2010 , pasp , 122 , 1035 asplund m. , grevesse n. , sauval a. j. & scott p. 2009 , ara&a , 47 , 481 andr p , ward - thompson d. & greaves j. 2012 , science , 337 , 69 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 1997 , a&a , 327 , 1054 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 1998 , a&a , 337 , 403 barber r. j. , tennyson j. , harris g. j. & tolchenov r. n. 2006 , mnras , 368 , 1087 basu s. & vorobyov e. i. 2012 , apj , 750 , 30 beer m. e. , king a. r. & pringle j. e. 2004 , mnras , 355 , 1244 bedin l. r. , anderson j. , king i. r. , piotto g. 2001 , apj , 560l , 75 bedin l. r. , giampaolo p. , king i. r. & anderson j. 2003 , aj , 126 , 247 bedin l. r. , salaris m. , piotto g. , anderson j. , king i. r. , cassisi s. 2009 , apj , 697 , 965 boudreault s. & lodieu n. 2013 , mnras , 434 , 142 braga v. f. , dallora m. , bono g. , stetson p. b. , ferraro i. et al . 2015 , apj , 799 , 165 burgasser a. j. , kirkpatrick j. d. , burrows a. , liebert j. , reid i. n. , gizis j. e. , mcgovern m. r. , prato l. , mclean i. s. 2003 apj , 592 , 1186 burgasser a. j. & kirkpatrick j. d. 2006 , apj , 645 , 1485 burgasser a. j. , witte s. , helling c. , sanderson r. e. , bochanski j. j. , hauschildt p. h. 2009 , apj , 697 , 148 burningham b. , smith l. , cardoso c. v. , lucas p. w. , burgasser a. j. , jones h. r. a. , smart r. l. , 2014 , mnras , 440 , 359 burrows a. , hubbard w. b. , saumon d. & lunine j. i. 1993 , apj , 406 , 158 burrows a. , marley m. hubbard w. b. et al . 1997 , apj , 491 , 856 burrows a. , hubbard w. b. , lunine j. i. & liebert j. 2001 , reviews of modern physics , 73 , 719 caffau e. , maiorca e. , bonifacio p. , faraggiana r. , steffen m. et . 2009 , a&a , 498 , 877 casewell s. l. , littlefair s. p. , burleigh m. r. & roy m. 2014 , mnras , 441 , 2644 chabrier g. , baraffe i. , allard f. & hauschildt p. h. 2000 , apj , 542 , 464 cushing m. c. , looper d. , burgasser a. j. , kirkpatrick j. d. , faherty j. et al . 2009 , apj , 696 , 986 dolphin 2000 , pasp , 112 , 1383 dotter a. , chaboyer b. , jevremovic d. , kostov v. , baron e. , ferguson j. w. 2008 , apjs , 178 , 89 elmegreen b. g. apj , 1999 , 522 , 915 fischer d. a. & valenti j. 2005 , apj , 622 , 1102 gnsicke b. , beuermann k. & demartino d. 1995 , a&a , 303 , 127 hansen b. m. s. , richer h. b. , fahlman g. g. , stetson p. b. , brewer j. et al . 2004 , apjs , 155 , 551 harris w.e . 1996 , aj , 112 , 1487 hasegawa , y. & hirashita , h. 2014 , apj , 788 , 62 hayashi , c. & nakano , t. 1963 , progress of theoretical physics , 30 , 460 hendricks b. , stetson p. b. , vandenberg d. a. & dallora m. 2012 , aj , 144 , 25 kaplan m. , stramatellos d. & whitworth a. p. 2012 341 , 395 kirkpatrick j. d. , schneider a. , fajardo - acosta s. , gelino c. r. , mace g. n. et al . 2014 , apj , 783 , 122 krist j. e. , hook r. n. & stoehr f. 2011 , proceedings of the spie , volume 8127 , i d . 81270j kroupa p. & bouvier j. 2003 , mnras , 346 , 369 kumar s. s. 1963 , apj , 137 , 1121 lpine s. , shara m. m. & rich r. m. 2004 , apjl , 602 , 125 lawrence a. , warren s. j. , almaini o. , edge a. c. , hambly n. c. et al . 2007 , mnras , 379 , 1599 luhman k. l. & sheppard s. s. 2014 , apj , 787 , 126 mace g. n , kirkpatrick j. d. , cushing m. c. , gelino c. r. , mclean i. s. et al . 2013 , apj , 777 , 36 malavolta l. , sneden c. , piotto g. , milone a. p. , bedin l. r. & nascimbeni v. 2014 , apj , 147 , 25 martn e. l. , dougados c. , magnier e. mnard f. , magazz a. , cuillandre j .- c . , delfosse x. 2001 , apj , 561 , 195 milone a. p. , piotto g. , bedin l. r. , aparicio a. , anderson j. et al . 2012 , a&a , 540 , 16 milone a. p. , marino a. f. , bedin l. r. , piotto g. , cassisi s. , dieball a. et al . 2014 , mnras , 439 , 1588 murray d. n. , burningham b. , jones h. r. a. , pinfield d. j. , lucas p. w. et al . 2011 , mnras , 414 , 575 pinfield d. j. , dobbie p. d. , jameson r. f. , steele i. a. , jones h. r. a. , katsiyannis a. c. 2003 , mnras , 342 , 1241 pinfield d. j. , gomes j. , day - jones a. c. , leggett s. k. , gromadzki m. et al . 2014 , mnras , 437 , 1009 rebolo r. , martin e. l. , basri g. , marcy g. w. & zapatero - osorio m. r. 1996 , apj , 469 , 53 reipurth b. & clarke c. 2001 , aj , 122 , 432 richer h. b. , fahlman g. g. , ibata r. a. et al . 1997 , apj , 484 , 741 richer h. b. , brewer j. , fahlman g. g. , kalirai j. , stetson p. b. , hansen b. m. s. , rich r. m. , ibata r. a. , gibson b. k. , shara , m. 2004 , aj , 127 , 2904 sarajedini a. , bedin l. r. , chaboyer b. , dotter a. , siegel m. et al . 2007 , aj , 133 , 1658 saumon d. & marley m. s. 2008 , apj , 689 , 1327 seaton , m. j. 1979 , mnras , 187 , 785 sigurdsson s. , richer h. b. , hansen b. m. , stairs i. h. , thorsett s. e. 2003 , science , 301 , 193 sivarani t. , lpine s. , kambhavi a. k. & gupchup j. 2009 , apjl , 694 , 140 skrutskie m. , cutri r. m. , stiening r. , weinberg m. d. , schneider s. et al . 2006 , aj , 131 , 1163 stamatellos d. , maury a. , whitworth a. & andr 2011 , mnras , 413 , 1787 stamatellos d. 2014 , assp , 36 , 17 steele i. a. , jameson r. f. , hodgkin s. t. & hambly n. c. 1995 . mnras , 275 , 841 thies i. & kroupa p. 2007 , apj , 671 , 767 thies i. , kroupa p. , goodwin s. , stamatellos d. , whitworth a. 2010 apj , 717 , 577 thies i. , pflamm - altenburg j. , kroupa p. & marks m. 2015 , apj accepted , arxiv:1501.01640 whitworth a. p. & goodwin , an , 326 , 899 wright e. l. , eisenhardt p. r. m. , mainzer a. k. , ressler m. e. , cutri r. m. et al . 2010 , aj , 140 , 1868 wood , m. a. 1995 , in white dwarfs , ed . d. koester & k. werner ( berlin : springer ) , 41 york d. g. , adelman j. , anderson j. e. , anderson s. f. , annis j. et al . 2000 , aj , 120 , 1579 zloczewski k. , kaluzny j. , rozyczka m. , krzeminski w. , mazur b. & thompson i. b. 2012 , acta astronomica , 62 , 357
we present an analysis of deep _ hst_/wfc3 near - ir ( nir ) imaging data of the globular cluster m4 . the best - photometry nir colour - magnitude diagram ( cmd ) clearly shows the main sequence extending towards the expected end of the hydrogen - burning limit and going beyond this point towards fainter sources . the white dwarf sequence can be identified . as such , this is the deepest nir cmd of a globular cluster to date . archival _ hst _ optical data were used for proper - motion cleaning of the cmd and for distinguishing the white dwarfs ( wds ) from brown dwarf ( bd ) candidates . detection limits in the nir are around @xmath0 mag and @xmath1 mag , and in the optical around @xmath2 mag . comparing our observed cmds with theoretical models , we conclude that we have reached beyond the h - burning limit in our nir cmd and are probably just above or around this limit in our optical - nir cmds . thus , any faint nir sources that have no optical counterpart are potential bd candidates , since the optical data are not deep enough to detect them . we visually inspected the positions of nir sources which are fainter than the h - burning limit in @xmath3 and for which the optical photometry did not return a counterpart . we found in total five sources for which we did not get an optical measurement . for four of these five sources , a faint optical counterpart could be visually identified , and an upper optical magnitude was estimated . based on these upper optical magnitude limits , we conclude that one source is likely a wd , one source could either be a wd or bd candidate , and the remaining two sources agree with being bd candidates . for only one source no optical counterpart could be detected , which makes this source a good bd candidate . we conclude that we found in total four good bd candidates .
recent submillimetre ( submm ) surveys show that the majority of the submm background at wavelengths of @xmath7 m arises from a population of distant , highly luminous infrared galaxies with 850@xmath0 m flux densities of @xmath8mjy ( blain et al . 1999 ; cowie et al . 2002 ; smail et al . the steep slope of the submm counts indicates that the integrated background is dominated by @xmath9mjy galaxies , below the @xmath10mjy confusion limit of the deepest blank - field 850@xmath0 m surveys carried out to date ( hughes et al.1998 ) , and well below the 38mjy flux density limits ( 34@xmath11 ) typically achieved in wider - field 850@xmath0 m surveys ( e.g. eales et al . 1999 ; scott et al . 2002 ; webb et al . 2003 ) . submm galaxies in the 1-mjy regime can , however , only be studied at present by exploiting a natural telescope a massive gravitational lens formed by a rich clusters of galaxies which provide the opportunity of overcoming both the confusion limit ( by spatially magnifying an area of the sky behind the lens ) and the sensitivity limit ( by gravitational magnification of the sources within this area ) . this strategy has successfully been used by smail , ivison & blain ( 1997 ) , chapman et al . ( 2002 ) , cowie et al . ( 2002 ) and smail et al . ( 2002 ) to detect submm sources amplified by factors typically of 1.54@xmath12 . unfortunately , the intrinsically faintest sources are usually only detected at modest significance , complicating the identification and analysis of their counterparts in other wavebands . to investigate the properties of the mjy - population in detail we require a high - signal - to - noise detection of a sub - mjy submm galaxy . such systems can arise in rare configurations where the source is located within a small region on the background sky defined by the caustic of the lens , which then forms multiple , highly magnified images of the background galaxy ( see also borys et al . this situation not only provides a tool for investigating the properties of intrinsically faint galaxies in detail , but also provides a unique opportunity to measure the redshift of the background galaxy from the geometry of the images combined with an accurate model of the cluster lens , using a ray - tracing technique to triangulate the three - dimensional position of the galaxy in the volume behind the lens . this technique , which was demonstrated successfully by kneib et al.(1996 ) and confirmed by ebbels et al . ( 1998 ) , circumvents the need for bright optical or infrared counterparts to a submm galaxy prior to measuring its redshift . we may thus avoid a bias in redshift measurements towards unrepresentative , low redshift or less obscured systems ( smail et al . 2002 ; webb et al . 2003 ) . this paper describes the discovery of a multiply - imaged submm galaxy , smmj16359 + 6612 , which is gravitationally lensed by the core of the rich cluster a2218 , and appears as three distinct sources in our 850@xmath0 m discovery map . we identify faint counterparts to all three sources in deep optical and near - infrared ( nir ) imaging , consistent with their identification as three images of a single background galaxy . we estimate its redshift using a detailed mass model for the cluster lens . we then confirm the accuracy of the redshift with spectroscopic observations and finally discuss the intrinsic properties of this system . throughout we will assume an @xmath13 , @xmath14 cosmology with @xmath15kms@xmath16mpc@xmath16 . at a redshift of @xmath5 the angular scale is thus 8.06 kpc / arcsec . see jpeg file a2218smm.jpg we used the submillimetre common user bolometer array ( holland et al.1999 ) on the 15-m james clerk maxwell telescope ( jcmt ) in march 1998 and in august 2000 to image a @xmath17 diameter field centred on the core of the @xmath18 cluster a2218 . total time on source excluding overheads was 95ks at a wavelength of 850@xmath0 m with chop - throws of 45 and 90@xmath19 at a variety of position angles . during the best observing conditions data were also collected at 450@xmath0 m ( 72ks in total without overheads ) . zenith opacities varied from 0.15 to 0.4 at 850@xmath0 m and from 0.6 to 1.0 at 450@xmath0 m . pointing was monitored and corrected every hour , and was found to be accurate to better than @xmath20 . the data were reduced using the standard surf package ( jenness & lightfoot 1998 ) . data reduction steps consisted of flat - fielding the demodulated data , opacity correction , despiking , flagging bad integrations , pointing correction , sky subtraction using the least noisy bolometers , and rebinning onto a @xmath21 grid , weighting the bolometers with the inverse variances of their signals . flux calibration was achieved using maps of uranus , and absolute flux calibration should be accurate to approximately 10% at 850@xmath0 m and 20% at 450@xmath0 m , as derived from the consistency of the calibration measurements . the resulting 850@xmath0 m image was convolved with a @xmath22 gaussian to remove high - frequency noise and subsequently deconvolved using a clean algorithm ( hogbom 1974 ) , in order to remove the negative sidelobes resulting from the small chop throw that was used for these data . after restoration with a @xmath23 full width at half maximum ( fwhm ) gaussian beam the resulting image has an r.m.s . noise of approximately 1.5mjybeam@xmath16 . [ cols="^,>,>,^,^,^,^,^ , > , > , > , > " , ] the optical images come from the _ hubble space telescope _ ( _ hst _ ) wfpc2 early release observations made after the sm-3a servicing mission in january 2000 . the dataset consists of 6 orbits in both f450w ( 11.0ks ) and f814w ( 12.0ks ) and 5 orbits in f606w ( 10.0ks ) . the reduction of these images is described in smail et al . the final wfpc2 frames have a fwhm of @xmath24 and 3-@xmath11 point source sensitivities of @xmath2528.8 , @xmath2629.0 and @xmath2728.1 in the vega - based system from holtzman et al . ( 1995 ) . the nir imaging used in our analysis was obtained with the ingrid camera ( packham et al . 2003 ) on the 4.2-m william herschel telescope ( wht ) . the observations were obtained during ingrid commissioning on march 22 and 23 , 2000 . the dataset consists of a total of 9.1ks integration in @xmath28 and 8.3ks in @xmath29-band obtained under photometric conditions in 0.60.8@xmath19 seeing . more details on the reduction of these data is described by smail et al . the absolute calibration is accurate to approximately 0.03mag . the final coadded frames have 3-@xmath11 point source sensitivities , within the seeing disk , of @xmath3022.0 and @xmath3123.7 , and seeing of @xmath32 . we overlay the 850@xmath0 m image on a true colour representation of the cluster core in [ fig.truecol ] . both the 850 and 450@xmath0 m scuba images show a number of bright submm sources located in the saddle region in the cluster core , formed by the two brightest cluster members . at 850@xmath0 m , at least four sources are readily identified above the 5-@xmath11 detection limit of @xmath33mjy . from north to south these are smmj16359 + 6613 ( @xmath34 , @xmath35 ) , smmj16359 + 6612.6 , smmj16359 + 6612.4 and smmj16358 + 6612.1 ( table 1 ) . the latter three sources also appear in the 450@xmath0 m map , but the northern source is not detected at 450@xmath0 m . significantly , the 450/850@xmath0 m flux ratios for the three southern sources are identical within the errors , @xmath36 . in contrast , the northern source has a 3-@xmath11 limit of @xmath37 . inspecting the positions of the four submm sources in the optical and nir data , we identify unusually red objects close to the position of all three southern submm sources , while the field of the northern submm source , smmj16359 + 6613 , shows a faint red counterpart visible in the _ hst / f814w _ frames but undetected in the ground - based @xmath28-band image . we show the relevant regions of the ingrid @xmath28 and _ hst _ frames in figure [ fig.truecol ] . it is worth noting that for all three southern sources the relative astrometric offsets between the @xmath28-band counterpart and the 850@xmath0 m peak is less than 2 . we label the optical / nir counterparts to the three southern submm sources from north to south : smm - a ( smmj16359 + 6612.6 ) , smm - b ( smmj16359 + 6612.4 ) , and smm - c ( smmj16358 + 6612.1 ) and list the photometry for these within matched apertures from our optical and nir frames in table [ tab.results ] . the colours of the three counterparts are similar , although crowding in the various fields and the effects of differential amplification on our fixed - aperture photometry may explain the lack of precise agreement ( the colours of smm - a should give the best indication of the true spectral energy distribution of the galaxy as it is not contaminated by any foreground galaxy ) . nevertheless , all three counterparts show a similar pattern of a pronounced red core surround by two blue regions , for example in smm - b the centre of the galaxy ( @xmath38 ) is much brighter in @xmath28 than in @xmath39 , while the southern ( and northern ) extremities , @xmath40 and @xmath41 , are considerably brighter in @xmath39 ( figure [ fig.truecol ] ) . a similar pattern of a red centre bracketed by bluer regions is seen in smm - a and c. we illustrate that this effect is not a result of the different spatial resolutions of the @xmath28 and @xmath39 frames by contouring the seeing - matched @xmath39 over the @xmath28-band frame in figure [ fig.truecol ] . the photometric similarity of the three southern submm sources , combined with the fact that their counterparts exhibit strong spectral and morphological similarities , suggests that we are seeing three images of a single luminous submm galaxy ( which we will call smmj16359 + 6612 in the following ) , lying behind the cluster . the configuration , relative brightnesses and parities of the three images are consistent with the predictions of a detailed mass model of the cluster core ( kneib et al . 1996 ) . with the assumption that these submm sources are three multiple images of a unique galaxy , we can predict a redshift of @xmath42 using the lens model of kneib et al . ( 1996 ) updated to include a recently discovered @xmath43 lensed galaxy ( ellis et al . 2001 ) . for comparison , the 450/850@xmath0 m flux ratio for this galaxy corresponds to a redshift / dust temperature value @xmath44 ( blain et al . 2002 ) or a redshift range of @xmath454.5 from the spectral energy distributions models in hughes et al . ( 1998 ) . on june 30 and july 1st 2003 , we conducted deep multi - slit spectroscopy with the low resolution imaging spectrograph ( lris ; oke et al . 1995 ) , of sources lying in the field of the rich cluster a2218 ( see kneib et al . 2004 for a complete description of these data and a detailled discussion of the mass model ) . the two nights had reasonable seeing , @xmath46 , but were not fully photometric ( with some cirrus ) , nevertheless we obtained a crude flux calibration of our observations using feige 67 and 110 as spectrophotometric standard stars . we observed the smm - a source for a total of 10.6ks using the 600/4000 blue grism and the 400/8500 red gratings offering respectively a spectral dispersion of 0.63 pixel@xmath16 in the blue and 1.86pixel@xmath16 in the red . we included in the same mask a slitlet on the galaxy # 468 ( the counter image of the @xmath47 fold arc # 384 : ebbels et al . ( 1996 ) , see also le borgne et al . ( 1992 ) ) . the spectra of smm - a and # 468 both show strong ly-@xmath40 absorption and numerous metal lines ( figure [ fig.spectra ] ) . the derived redshift for smm - a is @xmath48 based on the strongest interstellar absorption lines : [email protected] and [email protected] as well as [email protected] and 1550.77 in absorption . these observations also yielded a redshift for # 468 of @xmath50 . in addition , a 9.2-ks exposure using the 400/3400 blue grism on a second mask targeting galaxies in this field yielded a redshift for another @xmath51 galaxy . this galaxy is identified as # 273 ( le borgne et al . 1992 ) and was misidentified as a @xmath52 galaxy by ebbels et al . ( 1998 ) who confused the ciii]@xmath491908.7emission line with [ oii]@xmath493727 . as for smm - a and # 384/#468 , # 273 shows prominent ly-@xmath40 absorption and many metal absorption lines from which we derived a redshift of @xmath53 ( figure 2 ) . all three galaxies are thus within 100kms@xmath16 of each other in the rest frame ( ignoring uncertainties due to possible velocity offsets in the uv lines we use ) . figure [ fig.spectra ] , shows the uv spectra of all three galaxies with the identified absorption lines . the counterparts of the submm sources smm - a and smm - b were also independently targeted in the @xmath54-band using the nir spectrograph nirspec on keck ii ( mclean et al . these observations were obtained in 2002 august 18 in non - photometric conditions and 0.6 seeing . eight exposures of 600s each were obtained , nodding along the 0.76@xmath1242 slit by 8@xmath19 between each exposure . sources a and b were on the slit simultaneously . wavelength calibration was achieved from the sky lines , using observations of a bright star to correct for geometrical distortion . an average sky spectrum was created by scaling and combining the individual spectra , and this was then subtracted from each science observation before co - adding and extracting the spectra . approximate flux calibration and telluric absorption correction was achieved using an observation of ukirt standard stars from the previous night . no attempt was made to correct for slit losses . the flux - calibrated spectrum ( figure [ fig.spectra ] ) reveals three emission lines identified as h@xmath40 and the two weaker flanking [ nii ] lines . the h@xmath40 line extends over 2.2 for smm - a and 5.1 for smm - b , which compares well with the optical size of @xmath40 in smm - a and the total extent of @xmath55 and @xmath41 for smm - b as shown in the _ hst _ panels in figure [ fig.truecol ] and the two insets in figure [ fig.spectra ] . there are hints in smm - a for a more compact distribution of [ nii ] emission . the h@xmath40 line is centered on 2.30755@xmath0 m placing the submm galaxy at a redshift of @xmath56 , indicating that the uv interstellar lines used to derive a redshift in the optical are blueshifted by @xmath6kms@xmath16 with respect to the systemic velocity of the galaxy . the restframe width of the h@xmath40 line corresponds to @xmath57kms@xmath16 , and there is a hint of coherent velocity structure in the spectrum of smm - b , with a velocity shift of about @xmath58kms@xmath16 over an angular distance of 2.6 ( @xmath59kpc in the source plane ) . the fact that we do not see such structure in the smm - a image is due to the fact that only component @xmath40 ( and probably part of @xmath38 ) was in the nirspec slit . hence , this velocity shift could argue that the complex optical / nir morphology of this source is best explained by a strong dust absorption in a rotating disk structure or better by a merger . the total h@xmath40 line flux for the combined spectra of smm - a and b is @xmath60wm@xmath61 , with a ratio of 2.3 between their respective h@xmath40 flux densities . this corresponds to a h@xmath40 luminosity @xmath62l@xmath63 , or a ( unlensed ) star - formation rate of 11m@xmath63yr@xmath16 . the narrow width of the h@xmath40 emission line and the observed ratio of nii / h@xmath40 ( @xmath64 ) suggest that the emission is dominated by star formation , with little evidence for a strong agn . the occurrence of a submm galaxy falling within the caustic lines of a massive foreground gravitational lens provides unique constraints on the properties of this faint submm galaxy . using the cluster mass model we estimate that the background galaxy is gravitationally amplified by a factor of @xmath145 ( integrated across all three images ) , indicating that the intrinsic 850@xmath0 m flux density of this galaxy would be 0.8mjy in the absence of gravitational magnification , while in the optical and nir the galaxy would have magnitudes @xmath65 and @xmath2 . the galaxy therefore represents a serendipitously positioned example of the submm galaxy population at flux levels of @xmath9mjy , the population which produces the bulk of the submm background ( blain et al . this provides a unique opportunity to compare the properties of this low - luminosity submm galaxy with those of more luminous , submm galaxies studied in brighter , blank - field surveys ( e.g. ivison et al . 2002 ; chapman et al . 2003 ) . a second , multiply - imaged submm galaxy has been identified by borys et al . ( 2004 ) , giving two high - magnification examples from the 24 cluster lenses at @xmath66 mapped with sufficient sensitivity using scuba ( chapman et al . 2002 ; smail et al . 2002 ; knudsen et al . in prep . the surface density of @xmath67mjy submm galaxies is around 3arcmin@xmath61 ( smail et al . 2002 ; cowie et al . 2002 ) , suggesting that roughly ten clusters need to be surveyed to detect a strongly - lensed mjy - flux submm galaxy . this is consistent with the expectation based on models for the cluster lenses of uniform @xmath1800kms@xmath16 potential wells at @xmath68 , each with a critical curve encircling 0.05arcmin@xmath61 , and background submm galaxies at @xmath69 ( see kraiberg et al 2004 for a thorough discussion ) . in our assumed cosmology the far - infrared luminosity of smmj16359 + 6612 is @xmath70l@xmath63 , roughly 1.5@xmath12 fainter than arp220 , close to the border - line between luminous infrared galaxies and ultraluminous infrared galaxies ( ulirgs ) . the star formation rate derived from the far - infrared luminosity is about 500m@xmath63yr@xmath16 . figure 3 compares the restframe ( lens - corrected ) sed of the smmj16359 + 6612 to the seds of arp220 , the well - studied @xmath71 submm galaxy smmj02399@xmath720136 and the @xmath73 extremely red object and submm galaxy hr10 ( dey et al . the l@xmath74/l@xmath75 ratio is higher than that seen for arp220 , suggesting it is more obscured , and similar to the high - redshift , but much more luminous , smmj02399@xmath720136 ( ivison et al . 1998 ) . fig . 4 . the observed spectral energy distribution ( sed ) of the multiply - imaged submm galaxy smmj16359 + 6612 ( solid points with error bars ) compared to arp220 ( solid line , composite spectrum from anantharamaiah et al . 2000 , klaas et al . 1997 , lisenfeld et al . 1996 and surace et al . 2000 ) which is 1.5@xmath12 more luminous in the far - infrared , as well as the well - studied submm galaxy smmj02399@xmath720136 ( open diamonds , at @xmath71 , ivison et al . 1998 ) which is 10@xmath12 more luminous in the far - infrared , and the extremely red object hr10 ( open triangles , at @xmath73 , dey et al . 1999 ) which is 7@xmath12 more luminous in the far - infrared . all seds are redshifted to match that of smmj16359 + 6612 and scaled so they have a similar far - infrared luminosity . notice the relatively large variation of l@xmath74/l@xmath75 between the different galaxies . + [ fig.sed ] at @xmath47 the spatial resolution of the _ hst _ images , corresponds to @xmath76kpc , taking into account the gravitational magnification . thus the colour gradient within the background galaxy is on 12kpc scales similar to the obscured region in nuclear starbursts in ultraluminous infrared galaxies at the present - day . a constraint on the size of the submm emission region is obtained by subtracting point sources from the 850 and 450@xmath0 m maps . this procedure leaves no detectable residual of extended emission , indicating that to our measurement accuracy we are dealing with point sources . the tightest constraint is obtained at 450@xmath0 m where we find that the intrinsic extent of the emission is less than @xmath77 , which corresponds to @xmath78kpc . there is thus no indication for a highly extended obscured starburst in this relatively low - luminosity submm galaxy . the precise redshift we have measured for smmj16359 + 6612 enables us to identify that this galaxy lies at an identical redshift to that of # 273 and the # 384/#468 multiple image system . using the lens model we estimate that smmj16359 + 6612 lies between the # 384/#468 and # 273 galaxies in the source plane , and all three galaxies are less than 130kpc apart . if these galaxies were not magnified by the cluster lens , all three galaxies would appear within a radius of 8@xmath19 . moreover , we can place a limit of @xmath79kms@xmath16 on the possible velocity offset between these three galaxies . they are thus all part of a single group and it is likely that they are interacting , which may explain the activity we detect in the submm waveband . the strong clustering of the submm galaxies with uv - bright populations highlights the opportunity for measuring the distances to submm galaxies from the redshifts of less - obscured companions , as well as the possible confusion which may arise when trying to relate uv- and submm - selected populations , in the absence of precise positions from radio counterparts . smmj16359 + 6612 has much redder colour in @xmath80 and @xmath81 than the two other nearby uv - selected galaxies , confirming the dustier nature of this galaxy . if we compare the unlensed star formation rate ( sfr ) for this three systems , we find based on their uv continuum that : smmj16359 + 6612 has 6 m@xmath63yr@xmath16 ( although certainly underestimated due to dust extinction ) # 468 : 14 m@xmath63yr@xmath16 and # 273 : 4 m@xmath63yr@xmath16 . for smmj16359 + 6612 we can compare the three estimates of its amplification - corrected star formation rate : 6 m@xmath63yr@xmath16 from its dust - corrected uv luminosity , 11m@xmath63yr@xmath16 from the h@xmath40 flux ( uncorrected for extinction or aperture losses ) and 500m@xmath63yr@xmath16 based on the far - infrared emission . applying the median extinction correction derived for mid - ir selected luminous infrared galaxies at @xmath82 by flores et al . ( 2003 ) , a@xmath83 , and a modest correction for slit losses would increase the h@xmath40-derived star formation rate by a factor of @xmath84 . these results suggest the vast majority of the young stars in this galaxy are obscured by dust and are undetectable in the restframe far - uv . although it is 10@xmath12 less luminous than the typical submm - selected galaxy studied in blank - field surveys , this system shares the same extreme levels of obscuration seen in the more luminous galaxies , rather than the more modest dust obscuration inferred for the somewhat less luminous uv - selected galaxies at these redshifts . we also note that the h@xmath40 line width and velocity structure , if they reflect the dynamics of the galaxy , suggest that this system is more massive for its uv luminosity than typical uv - selected galaxies at this epoch ( erb et al . 2003 ) . in summary , we have identified a multiply - imaged submm galaxy seen through the core of the rich cluster a2218 . the cluster lens amplifies the background galaxy by a factor of @xmath85 , providing a high signal - to - noise view of an example of the sub - mjy submm population which provides the bulk of the extragalactic background in this waveband . we estimate a redshift for this galaxy from our highly - constrained lens models for the cluster and confirm this using optical and near - infrared spectroscopy from keck . the redshift of the submm galaxy is @xmath5 , placing it at the same redshift as two other strongly - lensed uv - bright galaxies in this field . our keck spectroscopy suggests that the emission from the submm galaxy is dominated by star formation and in contrast to the typical uv spectral properties of more luminous submm galaxies , this galaxy shows ly-@xmath40 absorption , rather than emission ( c.f . chapman et al.2003 ) . the star formation rates we derive from the uv continuum , h@xmath40 and far - infared emission show that most of the star formation activity is obscured and we suggest it is likely to be located in the component @xmath38 , given its extreme @xmath81 colour . because of its faint submm flux this galaxy is likely to be a good example of the type of galaxy that makes most contribution to the star formation history , thus deserving a detailled study with submm / mm interferometer to study the dynamics and mass of this obscured region . the presence of three highly - amplified @xmath47 galaxies in our survey field indicates that the submm galaxy resides in a compact group , interactions within which may help explaining the triggering of the obscured starburst we detect . we are grateful to r. blandford , a. fruchter , r. hook , c. packham , j. peacock , n. reddy , g. smith , r. tilanus and t. treu for their respective roles in the acquisition and reduction steps for the data that this paper is based upon . we thank the anonymous referee for his constructive and useful report . jpk acknowledges support from cnrs and caltech . irs acknowledges support from the royal society and the leverhulme trust . awb acknowledges support from the nsf under grant number ast-0205937 . we also acknowledge support from the uk french alliance collaboration programme # 00161xm . the jcmt is operated by the joint astronomy centre on behalf of the united kingdom particle physics and astronomy research council ( pparc ) , the netherlands organization for scientific research , and the national research council of canada . some of this data was obtained at the w. m. keck observatory , operated as a scientific partnership among caltech , the university of california and nasa . the observatory was made possible by the generous financial support of the w. m. keck foundation . the william herschel telescope is operated on the island of la palma by the isaac newton group in the spanish observatorio del roque de los muchachos of the instituto de astrofisica de canarias . this paper is based on observations obtained with the nasa / esa _ hubble space telescope _ , obtained from the data archive of the space telescope science institute which is operated by the association of universities for research in astronomy , inc . , under nasa contract nas5 - 26555 .
we present observations of a remarkable submillimetre - selected galaxy , smmj16359 + 6612 . this distant galaxy lies behind the core of a massive cluster of galaxies , a2218 , and is gravitationally lensed by the foreground cluster into three discrete images which were identified in deep submillimetre maps of the cluster core at both 450 and 850@xmath0 m . subsequent follow - up using deep optical and near - infrared images identify a faint counterpart to each of the three images , with similar red optical near - infrared colours and _ hst _ morphologies . by exploiting a detailed mass model for the cluster lens we estimate that the combined images of this galaxy are magnified by a factor of @xmath145 , implying that this galaxy would have un - lensed magnitudes @xmath2 and @xmath3 , and an un - lensed 850-@xmath0um flux density of only 0.8mjy . moreover , the highly constrained lens model predicted the redshift of smmj16359 + 6612 to be @xmath4 . we confirm this estimate using deep optical and near - infrared keck spectroscopy , measuring a redshift of @xmath5 . smmj16359 + 6612 is the faintest submm - selected galaxy so far identified with a precise redshift . thanks to the large gravitational magnification of this source , we identify three sub - components in this submm galaxy , which are also seen in the nirspec data , arguing for either a strong dust ( lane ) absorption or a merger . interestingly , there are two other highly - amplified galaxies at almost identical redshifts in this field ( although neither is a strong submm emitter ) . the three galaxies lie within a @xmath6kpc region on the background sky , suggesting this submm galaxy is located in a dense high - redshift group . [ firstpage ] galaxies : individual : smmj16359 + 6612 galaxies : starburst galaxies : evolution galaxies : clusters : individual : a2218 infrared : galaxies cosmology : observations
the discrete painlev equations , together with the painlev equations , are now widely recognized as one of the most important families of the integrable systems ( see , for example , @xcite ) . originally , the discrete painlev equations were discovered in the form of single second - order difference equations and identified as the discrete analogues of the painlev equations@xcite . then they were generalized to simultaneous first - order equations by a singularity confinement criterion@xcite . a typical example is the following equation known as a discrete painlev ii equation@xcite : @xmath3 where @xmath4 is the dependent variable , @xmath5 is the independent variable , and @xmath6 , @xmath7 , @xmath8 @xmath9 are parameters . we note that was found as the equation satisfied by the constant terms of the orthogonal polynomials on the unit circle@xcite . by applying the singularity confinement criterion , is generalized to @xmath10 where @xmath11 is a parameter , with its integrability preserved . introducing the dependent variables @xmath12 and @xmath13 by @xmath14 then can be rewritten as @xmath15 equation is known as a discrete painlev iii equation since it admits a continuous limit to the painlev iii equation@xcite . conversely , can be recovered from by putting @xmath16 and . this procedure is referred to as `` _ symmetrization _ '' of , which comes from the terminology of the quispel roberts thompson ( qrt ) mapping@xcite . after this terminology , is sometimes called the `` _ asymmetric _ '' discrete painlev ii equation , and is called the `` _ symmetric _ '' discrete painlev iii equation@xcite . it appears as though the symmetrization is a simple specialization on the level of the equation , but the following problems were known : ( i ) according to sakai s theory@xcite , the painlev and discrete painlev equations are classified by the underlying space of initial conditions . moreover , the discrete painlev equations arise as the birational mappings corresponding to the translations of the affine weyl groups associated with the space of initial conditions . the asymmetric discrete painlev equations are characterized in this manner , however , it was not known how to characterize the symmetric equations as the action of affine weyl groups . ( ii ) the painlev and discrete painlev equations admit the particular solutions expressible in terms of the hypergometric type functions ( hypergeometric solutions ) when some of the parameters take special values ( see , for example , @xcite and references therein ) . however , the hypergeometric solutions to the symmetric discrete painlev equation can not be obtained by the nave specialization of those to the corresponding asymmetric equation . for example , has the hypergeometric solution expressible in terms of the parabolic cylinder function ( weber function ) @xcite . on the other hand , admits the hypergeometric solution in terms of the confluent hypergeometric function@xcite . the crucial point is that although the former function is expressed as a specialization of the latter , this specialization is not consistent with the symmetrization . in @xcite , the mechanism of the symmetrization was investigated in detail by taking an example of @xmath0-painlev equation with the affine weyl group symmetry of type @xmath17 . then it was shown that in general , various discrete dynamical systems of painlev type can be obtained from elements of infinite order that are not necessarily translations in the affine weyl group by taking the projection on appropriate subspaces of the parameter spaces . such a procedure is called a _ projective reduction _ , and the symmetrization can be understood as a kind of projective reduction . moreover , the above nontrivial inconsistency among the hypergeometric solutions are explained by the factorization of the linear difference operators associated with the three - term relation of the hypergeometric functions . in spite of understanding the mechanism of the symmetrization or the projective reduction , it is still nontrivial what function will appear in the hypergeometric solutions to the symmetric discrete painlev equations even if the hypergeometric solutions to the corresponding asymmetric equations are known . since the continuous limit of the symmetric discrete painlev equation is different from the corresponding asymmetric one , it is also important to consider the continuous limit of the symmetric discrete painlev equations together with their hypergeometric solutions . the purpose of this paper is to construct the simplest hypergeometric solutions to each of the symmetric @xmath0-painlev equations . in @xcite the simplest hypergeometric solutions to all possible @xmath0-painlev equations in sakai s list have been constructed . similarly , in this paper , we present the list of the simplest hypergeometric solutions to the symmetric @xmath0-painlev equations reduced from the asymmetric ones with the affine weyl group symmetry of type @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 and @xmath17 . we also aim to consider the continuous limits of those symmetric @xmath0-painlev equations to the painlev equations , together with their hypergeometric solutions . this paper is organized as follows : in section 2 , we derive the symmetric @xmath0-painlev equations by applying the projective reduction to @xmath0-p@xmath23 , @xmath0-p@xmath24 , @xmath0-p@xmath25 , @xmath0-p@xmath26 , @xmath0-p@xmath27 and @xmath0-p@xmath28 , where the @xmath0-painlev equation with the affine weyl group of type @xmath29 is denoted by @xmath0-p(@xmath29 ) . in section 3 , we construct the hypergeometric solutions to the symmetric @xmath0-painlev equations derived in section 2 . in section 4 , we discuss the continuous limits of the symmetric @xmath0-painlev equations and their hypergeometric solutions . in section 5 , we prove that hypergeometric function appearing in the hypergeometric solution to the symmetric @xmath0-p@xmath24 actually reduces to the weber function by applying the saddle point method to its integral representation . some concluding remarks are given in section 6 . in this section , we apply the projective reduction to the @xmath0-painlev equations @xmath0-p@xmath23 , @xmath0-p@xmath24 , @xmath0-p@xmath25 , @xmath0-p@xmath26 , @xmath0-p@xmath27 and @xmath0-p@xmath28 , respectively , to obtain their symmetric forms . in the following , we use the notations @xmath30 @xmath0-p@xmath23 is given by @xcite @xmath31 the continuous limit yields the painlev iii equation . to apply the projective reduction , we put @xmath32 then is reduced to the symmetric @xmath0-p@xmath23 @xcite @xmath33 the continuous limit yields the painlev ii equation . the system described in is a discrete dynamical system arising from the birational action of an element @xmath34 of the affine weyl group of type @xmath35 , which is a translation on the corresponding root lattice and parameter space . we introduce another element of the affine weyl group @xmath36 satisfying @xmath37 . we note that @xmath36 is not a translation on the root lattice or on the full parameter space . however , by taking the projection on an appropriate subspace of the parameter space , it becomes a translation on the subspace , which gives . the specialization from to corresponds to this projection , which is an example of a projective reduction . we refer to @xcite for details . @xmath0-p(@xmath22)@xcite : @xmath38 symmetric @xmath0-p(@xmath22)@xcite : @xmath39 where @xmath40 the continuous limits of and yield the painlev v and iv equations , respectively . @xmath0-p(@xmath21)@xcite : @xmath41 symmetric @xmath0-p(@xmath21)@xcite : @xmath42 where @xmath43 the continuous limits of and yield the painlev vi and iii equations , respectively . @xmath0-p(@xmath20)@xcite : @xmath44 ( gf-1)(g\underline{f}-1)=q^{-1}t^2\cfrac{(g-{b_1}^{-1})(g-{b_2}^{-1})(g-{b_3}^{-1})(g-{b_4}^{-1})}{(g - b_6u)(g-{b_6}^{-1}u ) } , \end{cases}\ ] ] where @xmath45 symmetric @xmath0-p(@xmath20)@xcite : @xmath46 where @xmath47 direct continuous limit of to the painlev equations does not exist , since it has more parameters than the painlev vi equation . on the other hand , admits a continuous limit to the painlev v equation . @xmath0-p(@xmath19)@xcite : @xmath48 \cfrac{(gf - t^2)(g\underline{f}-t\underline{t})}{(gf-1)(g\underline{f}-1 ) } = \cfrac{(g-{b_1}^{-1}t)(g-{b_2}^{-1}t)(g-{b_3}^{-1}t)(g-{b_4}^{-1}t ) } { ( g-{b_5}^{-1})(g-{b_6}^{-1})(g-{b_7}^{-1})(g-{b_8}^{-1 } ) } , \end{cases}\ ] ] where @xmath49 symmetric @xmath0-p(@xmath19)@xcite : @xmath50 where @xmath51 direct continuous limit of to the painlev equations does not exist , while that of yields the painlev vi equation . @xmath0-p(@xmath18)@xcite : @xmath52 \cfrac{(fut - g)(\underline{f}u\underline{t}-g)-(u^2t^2 - 1)(u^2\underline{t}^2 - 1 ) } { ( u^{-1}t^{-1}f - g)(u^{-1}\underline{t}^{-1}\underline{f}-g)-(1-u^{-2}t^{-2})(1-u^{-2}\underline{t}^{-2 } ) } = \cfrac{p(g , u , m_7,\cdots , m_1)}{p(g , u^{-1},m_1,\cdots , m_7 ) } , \end{cases}\ ] ] where @xmath53 here @xmath54 @xmath55 are the elementary symmetric functions of @xmath56-th degree in @xmath57 @xmath58 , @xmath59 , and @xmath60 . symmetric @xmath0-p(@xmath18 ) : @xmath61 where @xmath62 direct continuous limits of and to the painlev equations are not known . in this section , we construct the hypergeometric solutions to the symmetric @xmath0-painlev equations . we use the following conventions of @xmath0-analysis@xcite . the basic hypergeometric series @xmath63 is defined by @xmath64 where @xmath65 are the @xmath0-shifted factorials . the following special case of the basic hypergeometric series is known as the very - well - poised basic hypergeometric series @xmath66 we also use the jacobi theta function @xmath67 which satisfies the @xmath0-difference equation @xmath68 [ prop : hyper_sym_a2a1 ] symmetric @xmath0-p@xmath23 admits the hypergeometric solution @xmath69 @xmath70 with @xmath71 here , @xmath72 and @xmath73 are quasi - constants satisfying @xmath74 and @xmath75 , respectively . * proof . * substituting @xmath76 in , we find that admits a specialization to the discrete riccati equation @xmath77 when @xmath78 . then putting as ) , is linearized to the three - term relation for @xmath79 @xmath80 moreover , substituting the power series expression @xmath81 into ) , we obtain .@xmath82 in the following , we present the hypergeometric solutions to other cases . since the results are verified by direct calculations , we omit the proof showing only the discrete riccati equations and the linearized three - term relations . we also assume that @xmath72 and @xmath73 are quasi - constants . [ prop : hyper_sym_a4 ] symmetric @xmath0-p@xmath24 admits the hypergeometric solution @xmath83 @xmath84 with @xmath85 the discrete riccati equation and the linearized three - term relation are given by @xmath86 and @xmath87 respectively . for convenience , we set @xmath88 symmetric @xmath0-p@xmath89 can be rewritten as @xmath90 [ prop : hyper_sym_d5 ] symmetric @xmath0-p@xmath25 admits the hypergeometric solution @xmath91 @xmath92 with @xmath93 here , @xmath94 is jackson s @xmath0-bessel fuction@xcite @xmath95 the discrete riccati equation and the linearized three - term relation are given by @xmath96 and @xmath97 respectively . [ prop : hyper_sym_e6 ] symmetric @xmath0-p@xmath26 admits the hypergeometric solution @xmath98 @xmath99 with @xmath100 the discrete riccati equation and the linearized three - term relation are given by @xmath101 @xmath102 respectively . [ prop : hyper_sym_e7 ] symmetric @xmath0-p@xmath27 admits the hypergeometric solution @xmath103 @xmath104 with @xmath105 the discrete riccati equation and the linearized three - term relation are given by @xmath106 @xmath107 respectively . to obtain the solution of , the following proposition is useful : [ prop:3term_askey ] the three - term relation for @xmath108 , where @xmath109 is given by @xmath110 here @xmath72 and @xmath73 are quasi - constants with respect to @xmath111 . substituting @xmath112 in proposition [ prop:3term_askey ] , we obtain the solution of . for later convenience , we rewrite the solution in the following manner : @xmath113 then proposition [ prop : hyper_sym_e7 ] is rephrased as follows : symmetric @xmath0-p@xmath27 admits the hypergeometric solution @xmath114,\ ] ] @xmath115 when @xmath116 we note that the linear three - term relation for @xmath117 is given by @xmath118 [ prop : hyper_sym_e8 ] symmetric @xmath0-p@xmath28 admits the hypergeometric solution @xmath119,\ ] ] @xmath120 with @xmath121 here , @xmath122 and @xmath123 are given by @xmath124 the discrete riccati equation and the linearized three - term relation are given by @xmath125 @xmath126 respectively . we note that is obtained by substituting @xmath127 in proposition [ prop:3term_askey ] . in this section , we discuss the continuous limits of the symmetric @xmath0-painlev equations and their hypergeometric solutions . we introduce the hypergeometric series @xmath128 defined by @xmath129 where @xmath130 . it is known that symmetric @xmath0-p@xmath23 yields the painlev ii equation @xmath131 where @xmath132 , by putting @xmath133 and taking the limit @xmath134 @xcite . the limit of the hypergeometric solution is given as follows : under the parametrization , the hypergeometric solution in proposition [ prop : hyper_sym_a2a1 ] is reduced to the following solution to : @xmath135 @xmath136 with @xmath137 here @xmath72 and @xmath73 are constants , and @xmath138 is the airy function @xmath139 we note that and are reduced to the following equations @xmath140 @xmath141 respectively . we also note that is obtained from by using the following proposition : [ prop : continuous_airy ] with the substitutions @xmath142 as @xmath143 , it follows that @xmath144,\\ & { } _ 1\varphi_1 \left(\begin{matrix}0\\-q\end{matrix};q ,- qt\right ) = 2\pi^{1/2}\delta^{-1/2}e^{-\log2(\pi i/\delta^3)+(\pi i/2\delta)u+\pi i/12 } \left[{\rm ai}(ue^{\pi i/3})+o(\delta^2)\right],\end{aligned}\ ] ] for @xmath111 in any compact domain of @xmath145 . putting @xmath146 and taking the limit @xmath147 , symmetric @xmath0-p@xmath24 yields the painlev iv equation @xmath148 [ prop : a4_sol ] under the parametrization , the hypergeometric solution in proposition [ prop : hyper_sym_a4 ] is reduced to @xmath149 @xmath150 with @xmath151 here @xmath72 and @xmath73 are constants , and @xmath152 is the weber function @xmath153 where the path of integration @xmath154denote it by @xmath155@xmath156 runs from @xmath157 to @xmath158 so that @xmath159 lies to the right of the path . the hypergeometric solution to the painlev iv equation given in proposition [ prop : a4_sol ] is consistent with the solution appeared in @xcite . note that and are reduced to the following equations @xmath160 @xmath161 respectively . we also note that is obtained from by using the following lemmas : [ lemma : continuous_weber_1 ] with the substitutions , it follows that @xmath162~du , \end{split}\ ] ] as @xmath163 the path @xmath164 runs from @xmath157 to @xmath158 when @xmath165 and @xmath158 to @xmath157 when @xmath166 so that @xmath159 lies to the right of the path . [ lemma : continuous_weber_2 ] with the substitutions , it follows that @xmath167du , \end{split}\ ] ] as @xmath168 the path @xmath164 runs from @xmath157 to @xmath158 when @xmath169 and @xmath158 to @xmath157 when @xmath170 so that @xmath159 lies to the right of the path . the proofs of lemma [ lemma : continuous_weber_1 ] and lemma [ lemma : continuous_weber_2 ] will be given in the next section . for these proofs , we use the saddle point method@xcite . moreover , for the asymptotic expansions of the @xmath0-shifted factorials as @xmath171 @xmath172 , we use the following proposition : [ prop : qshifted_asymptotic ] as @xmath171 , the @xmath0-shifted factorials have an asymptotic expansions , @xmath173 this is uniform for @xmath174 such that @xmath175 . here @xmath176 is the euler dilogarithm defined by @xmath177 putting @xmath178 and taking the limit @xmath179 , symmetric @xmath0-p@xmath25 yields the painlev iii equation@xcite @xmath180 under the parametrization , the hypergeometric solution in proposition [ prop : hyper_sym_d5 ] is reduced to @xmath181 @xmath182 with @xmath183 here , @xmath72 and @xmath73 are constants and @xmath184 is the bessel function @xmath185 we note that and are reduced to the following equations @xmath186 @xmath187 respectively . putting @xmath188 and taking the limit @xmath179 , symmetric @xmath0-p@xmath26 yields the painlev v equation@xcite @xmath189 [ prop : p5_sol ] under the parametrization , the hypergeometric solution in proposition [ prop : hyper_sym_e6 ] is reduced to the following solution to : @xmath190 @xmath191 with @xmath192 here , @xmath72 and @xmath193 are constants . the hypergeometric solution in proposition [ prop : hyper_sym_e6 ] is reduced to the following solution to @xmath194 @xmath195 respectively . we note that is obtained from by the following procedure . by using heine s transformation@xcite @xmath196 and setting @xmath197 is rewritten as @xmath198 then yields the hypergeometric series in by taking a term - by - term limit . we use proposition [ prop : qshifted_asymptotic ] for the limit of the coefficient . setting @xmath199 we can rewrite the riccati equation , the dependent variable teansformation and the linear differential equation as @xmath200 respectively . these coincide with the result in @xcite . we note here that in @xcite the solution to is given by the hypergeometric series around zero while in this section that is given by the series at infinity . putting @xmath201 and taking the limit @xmath179 , symmetric @xmath0-p@xmath27 yields painlev vi equation @xmath202 where @xmath203@xcite . under the parametrization , the hypergeometric solution in proposition [ prop : hyper_sym_e7 ] is reduced to the following solution to : @xmath204 @xmath205 with @xmath206 here , @xmath72 and @xmath193 are constants . we note that and are reduced to the following equations @xmath207 @xmath208 respectively . we also note that is obtained by the following procedure . by using the transformation ( see , ( iii.23 ) in @xcite ) @xmath209 and setting @xmath210 is rewritten as @xmath211 then is obtained from by taking a term - by - term limit and using proposition [ prop : qshifted_asymptotic ] . in this section , we give the proof of lemma [ lemma : continuous_weber_1 ] and [ lemma : continuous_weber_2 ] . taking the continuous limit term - by - term in the hypergeometric series in does not yield any meaningful result . in this case , we use the saddle point method@xcite . consider the complex integral @xmath212 where @xmath213 and @xmath214 are analytic functions . the point @xmath215 satisfying @xmath216 and the direction @xmath217 are called a saddle point and a steepest descent direction , respectively . when the path @xmath218 passes in the steepest descent direction , @xmath219 reaches its peak at the saddle point . therefore @xmath220 may be evaluated only around the saddle point when @xmath221 . for the asymptotic expansions of the @xmath0-shifted factorials as @xmath171 , we use proposition [ prop : qshifted_asymptotic ] . we here fix the branch of @xmath222 and fractional power functions as @xmath223 where @xmath224 . note that @xmath225 and @xmath226 are valid only @xmath227 . we consider the continuous limit of @xmath228 with @xmath229 we note that in taking the limit , the sign of @xmath230 is chosen according to the value of @xmath231 as shown later . by using heine s transformation@xcite @xmath232 @xmath233 is rewritten as @xmath234 we next prepare a suitable integral representation for @xmath233 . [ lemma:1phi1_integral ] it holds that @xmath235 where the path @xmath218 runs from @xmath236 to @xmath237 so that the poles of @xmath238 lie to the right of the path and the other poles lie to the left of the path @xmath154as shown in figure [ fig : contour_c]@xmath156 . * we derive according to the method in @xcite . we consider the integral @xmath239 where the contour @xmath240 is a large clockwise - oriented semicircle of radius @xmath241 with center at the origin and circle round only part of the poles of @xmath238 ( as shown in figure [ fig : contour_cn ] ) . by the residue theorem , we have @xmath242 therefore we obtain @xmath243 in order to estimate the contribution from @xmath244 we need a bound on the integrand for large @xmath245 @xmath246 as it is dominated by the product in the denominator . from @xmath247 and , we have completed the proof.@xmath82 , scaledwidth=100.0% ] , scaledwidth=100.0% ] by using lemma [ lemma:1phi1_integral ] , it holds that @xmath248 where the path @xmath249 runs from @xmath236 to @xmath237 so that the poles of @xmath250 lie to the right of the path and the other poles lie to the left of the path . we divide the integral path @xmath249 into @xmath251 which runs from @xmath236 to @xmath237 so that all poles lie to the right of the path and @xmath252 which is an anticlockwise - oriented contour and encircles all poles except for the poles of @xmath250 ( as shown in figure [ fig : contour_l1_hatc1 ] ) . we now evaluate @xmath233 at @xmath147 by using the saddle point method . and @xmath252,scaledwidth=50.0% ] * ( 1 ) path @xmath251.*set @xmath253 using proposition [ prop : qshifted_asymptotic ] , we rewrite @xmath254 as @xmath255~dz,\ ] ] where @xmath256 by the identity of dilogarithm@xcite @xmath257 @xmath258 is rewritten as @xmath259 let us first find a zero point of @xmath260 ( saddle point ) . differentiating by @xmath261 and using , we obtain @xmath262 we note here that only @xmath263 can be saddle points at @xmath221 . [ lemma : rouche ] there exists only one saddle point in the domain @xmath264 . * proof . * set @xmath265 functions @xmath260 and @xmath266 are regular in @xmath267 . furthermore , since @xmath268 it is obvious that @xmath269 on the circle @xmath270 . by rouch s theorem , the numbers of zeros of @xmath271 are equal to those of @xmath260 . therefore we have completed the proof . @xmath82 from lemma [ lemma : rouche ] , the saddle point @xmath272 is determined by assuming the series expansion @xmath273 in as @xmath274 since @xmath275 the steepest descent direction is given as @xmath276 in order that @xmath251 passes in the steepest descent direction , it is necessary from the configuration of the paths ( see figure [ fig : contour_l1_hatc1 ] ) that @xmath277 therefore we obtain the following condition @xmath278 now , we consider an approximation of the integrand of @xmath254 . since the integrand of @xmath254 is evaluated only around the saddle point @xmath279 , we assume that @xmath261 is in the set @xmath280 for simplicity , we first change the variable @xmath261 to @xmath281 so that the saddle point is given as @xmath282 . from , we set @xmath281 so as to satify @xmath283 moreover , from and to calculate the integration by substitution from @xmath261 to @xmath281 simply , we also require @xmath284 that is , @xmath285 we now add a condition between @xmath261 and @xmath281 so that the correspondence @xmath286 is @xmath287 : @xmath287 . to this end , we fix the branch at the point @xmath288 corresponding to @xmath289 . from @xmath290 satisfies @xmath291 we fix the branch by choosing @xmath292 , that is , from @xmath281 is expressed by @xmath261 as @xmath293 furthermore , @xmath261 can be also expressed by @xmath281 . noticing that @xmath294 , we obtain from @xmath295 since @xmath289 corresponds to @xmath296 , it holds that @xmath297 for later convenience , we consider the point @xmath298 corresponding to @xmath299 . by the assumption , @xmath300 is given in the form @xmath301 where @xmath302 and @xmath303 . from , we have @xmath304 comparing the degree of @xmath230 of , we obtain @xmath305 or @xmath306 . from , it holds that @xmath307 which indicates @xmath308 therefore we can determine @xmath300 from as @xmath309 we proceed to the approximation of integrand of @xmath254 by changing the variable from @xmath261 to @xmath281 . we expand @xmath310 in the form @xmath311 since the characteristic exponent of @xmath310 at @xmath312 is @xmath313 ( see ) , it holds that @xmath314 from , and , we obtain @xmath315d\hat{z}\\ = & e^{(\pi i\log2+\pi^2/2)/\epsilon^2-\pi s/\epsilon}c_0 \int_{l_1}e^{(\hat{z}^2 + 4is\epsilon \hat{z})/4\epsilon^2 } \hat{z}^{\alpha-1}[1+o(\epsilon)]d\hat{z}.\label{eqn : i1_c0}\end{aligned}\ ] ] in the derivation of , we note that since @xmath310 does not diverge at @xmath221 , @xmath316 also does not diverge . @xmath317 in can be determined in the following manner . from and , we heve @xmath318 substituting @xmath319 in , we get @xmath320 in order to adjust to the integral representation of the weber function , we introduce the variable @xmath111 by @xmath321 . then we obtain @xmath322 du\\ = & ( -1)^{3\alpha/2 + 1/2}2^{-\alpha/2}\epsilon^\alpha e^{(\pi i\log2+\pi^2/2)/\epsilon^2-\pi s/\epsilon } \int_{\hat{l}_1}e^{-u^2/2 - 2^{1/2}su}u^{\alpha-1}[1+o(\epsilon ) ] du,\label{eqn : value_i1}\end{aligned}\ ] ] where the path of integration @xmath323 runs from @xmath157 to @xmath324 when @xmath165 and @xmath324 to @xmath157 when @xmath166 so that @xmath159 lies to the right of the path . * ( 2 ) contour @xmath252.*set @xmath325 to evaluate @xmath326 , we prepare the following lemma : [ lemma:2phi1_integral_1 ] it holds that @xmath327 where @xmath218 is an anticlockwise - oriented contour and encircles all poles except for the poles of @xmath238 . * proof . * we consider the integral @xmath328 where @xmath329 is an anticlockwise - oriented contour and encircles the poles of @xmath330 . by the residue theorem , we obtain @xmath331 then the statement follows from the limit of @xmath332.@xmath82 by using lemma [ lemma:2phi1_integral_1 ] , @xmath326 is rewritten as @xmath333 finally by using heine s transformation@xcite @xmath334 we obtain @xmath335 from , and , we have @xmath336~du\notag\\ & + \cfrac{({a_2}^{-2};p)_\infty e^{\pi i\log ( p{a_2}^{-1})/\log p}}{(-{a_2}^2;p)_\infty(-1)^{\log t/\log p}}~ { } _ 2\varphi_1\left(\begin{matrix}-{a_2}^2,0\\-p\end{matrix};p , p{a_2}^{-1}t\right),\end{aligned}\ ] ] which proves lemma [ lemma : continuous_weber_1 ] . we consider the continuous limit of @xmath337 with . by using heine s transformation@xcite @xmath338 @xmath339 is rewritten as @xmath340 we prepare a suitable integral representation for @xmath339 . [ lemma:1phi1_integral_2 ] it holds that @xmath341 where the path @xmath218 runs from @xmath157 to @xmath158 so that the poles of @xmath238 lie to the right of the path and the other poles lie to the left of it . the lemma [ lemma:1phi1_integral_2 ] can be proved in a similar manner to lemma [ lemma:1phi1_integral ] . although lemma [ lemma:1phi1_integral ] and lemma [ lemma:1phi1_integral_2 ] differ only in the integration path , it is appropriate to apply lemma [ lemma:1phi1_integral_2 ] in order to choose the integration path in the steepest descent direction . by using lemma [ lemma:1phi1_integral_2 ] , it holds that @xmath342 where the path @xmath343 runs from @xmath157 to @xmath158 so that the poles of @xmath250 lie to the right of the path and the other poles lie to the left of it . we divide the path @xmath343 into @xmath344 which runs from @xmath157 to @xmath158 so that all poles lie to the right of the path and @xmath345 which is an anticlockwise - oriented contour and encircles all poles except for the poles of @xmath250 ( as shown in figure [ fig : contour_l2_hatc2 ] ) . and @xmath345,scaledwidth=90.0% ] * ( 1 ) path @xmath344 . * set @xmath346 in a similar manner to @xmath254 , one can show that @xmath347du,\ ] ] under the assumption @xmath348 here the path @xmath349 runs from @xmath157 to @xmath158 when @xmath169 and @xmath158 to @xmath157 when @xmath170 so that @xmath159 lies to the right of the path . * ( 2 ) contour @xmath345.*set @xmath350 in the following , we transform the integral in into another suitable integral , then evaluate its asymptotic behavior . more precisely , we first identify @xmath351 with appropriate basic hypergeometric function , then apply heine s transformation and reconstruct the integral representation of it . we next evaluate its asymptotic behavior . [ lemma:2phi1_integral_2 ] it holds that @xmath352 where @xmath218 is an anticlockwise - oriented contour and encircles all poles except for the poles of @xmath250 . lemma [ lemma:2phi1_integral_2 ] can be proved in a similar manner to lemma [ lemma:2phi1_integral_1 ] . then @xmath351 is rewritten by using lemma [ lemma:2phi1_integral_2 ] as @xmath353 further , by applying heine s transformation@xcite @xmath354 and using lemma [ lemma:1phi1_integral_2 ] , we obtain @xmath355 where the path @xmath356 runs from @xmath157 to @xmath158 so that the poles of @xmath250 lie to the right of the path and the other poles lie to the left of it . we finally evaluate the integral of the right hand side of ( 5.77 ) . to this end , we divide the path @xmath356 into @xmath357 which runs from @xmath157 to @xmath158 so that all poles lie to the right of the path and @xmath358 which is an anticlockwise - oriented contour and encircles all poles except for the poles of @xmath250 . set @xmath359 in a similar manner to @xmath254 , the asymptotic behavior of @xmath360 as @xmath361 can be evaluated as @xmath362~du.\ ] ] @xmath363 can be expressed in terms of the basic hypergeometric function . in fact , we have the following lemma which is proved in a similar manner to lemma [ lemma:2phi1_integral_1 ] : [ lemma:2phi1_integral_3 ] it holds that @xmath364 where @xmath218 is an anticlockwise - oriented contour and encircles all poles except for the poles of @xmath250 . then by using lemma [ lemma:2phi1_integral_3 ] , @xmath363 yields @xmath365 which is rewritten by applying heine s transformation as @xmath366 therefore we finally obtain @xmath367du\notag\\ & + \cfrac{(-{a_2}^2;p)_\infty e^{(2\log ( -pa_2)\log a_2-\log ( p{a_2}^{-1})\log ( -{a_2}^{-2}))/\log p } } { { a_2}^4({a_2}^2;p)_\infty}~(-1)^{\log t/\log p } { } _ 2\varphi_1\left(\begin{matrix}{a_2}^2,0\\-p\end{matrix};p , p{a_2}^{-1}t\right),\end{aligned}\ ] ] where we have used proposition [ prop : qshifted_asymptotic ] . this proves lemma [ lemma : continuous_weber_2 ] . in this paper , we have constructed the hypergeometric solutions to the symmetric @xmath0-painlev equations of the types @xmath17 , @xmath22 , @xmath21 , @xmath20 , @xmath19 and @xmath18 , and discussed their continuous limits . in particular , we have shown that the hypergeometric function appearing in the solution to the symmetric @xmath0-p@xmath24 actually reduces to the weber function by applying the saddle point method to its integral representation . before closing , we give a remark on a @xmath0-painlev equation of the type @xmath22 . the following @xmath0-difference equation @xmath368 is usually referred to as a @xmath0-painlev equation of type @xmath22 . equation also describes a translation on the root lattice of type @xmath22 , but its direction is different from that of . it may be interesting to note that the symmetric @xmath0-painlev equation obtained from by the projective reduction @xmath369 has no hypergeometric solution ( see @xcite ) which is a sharp contrast to . * acknowledgement . * the authors would like to express their sincere thanks to mr . p. howes , prof . h. ochiai and prof . t. tsuda for fruitful discussions and valuable suggestions . this work has been partially supported by jsps grant - in - aid for scientific research no . 23340037 and 22@xmath3704366 . 99 e. brzin and v.a . kazakov , exactly solvable theories of closed strings , phys . b * 236 * ( 1990 ) 144150 . douglas and s.h . shenker , strings in less than one dimension , nucl . b * 335 * ( 1990 ) 635654 . fokas , a.r . its and a.v . kitaev , the isomonodromy approach to matrix models in 2d quantum gravity , comm . * 147 * ( 1992 ) 395430 . b. grammaticos , f.w . nijhoff , v. papageorgiou , a. ramani and j. satsuma , linearization and solutions of the discrete painlev iii equation , phys . lett . a * 185 * ( 1994 ) 446452 . g. gasper and m. rahman , basic hypergeometric series ( second edition ) , encyclopedia of mathematics and its applications * 96 * ( cambridge university press , cambridge , 2004 ) . b. grammaticos and a. ramani , on a novel @xmath0-discrete analogue of the painlev vi equation , phys . lett . a * 257 * ( 1999 ) 288292 . b. grammaticos and a. ramani , discrete painlev equations : a review , lecture notes in physics * 644 * ( 2004 ) 245321 . b. grammaticos , a. ramani and v. papageorgiou , do integrable mappings have the painlev property ? . lett . * 67 * ( 1991 ) 18251828 . t. hamamoto , k. kajiwara and n.s . witte , hypergeometric solutions to the @xmath0-painlev equation of type @xmath371 , int . res . not . * 2006 * ( 2006 ) article i d 84619 . ismail and m. rahman , the associated askey - wilson polynomials , trans . * 328 * ( 1991 ) 201237 . m. jimbo and h. sakai , a @xmath0-analog of the sixth painlev equation , lett . * 38 * ( 1996 ) 145154 . kirillov , dilogarithm identities , progr . phys . suppl . * 118 * ( 1995 ) 61142 . k. kajiwara , the discrete painlev ii equation and the classical special functions , in _ symmetries and integrability of difference equations _ , eds . by p. clarkson and f.w . nijhoff , london math . lecture note ser . * 255 * ( cambridge university press , cambridge , 1999 ) 217227 . k. kajiwara , t. masuda , m. noumi , y. ohta and y. yamada , hypergeometric solutions to the @xmath0-painlev equations , int . res . not . * 2004 * ( 2004 ) 24972521 . k. kajiwara , t. masuda , m. noumi , y. ohta and y. yamada , construction of hypergeometric solutions to the @xmath0-painlev equations , int . res . not . * 2005 * ( 2005 ) 14411463 . k. kajiwara , n. nakazono and t. tsuda , projective reduction of the discrete painlev system of type @xmath17 , int . math . res . not . * 2011 * ( 2011 ) 930966 . k. kajiwara , y. ohta and j. satsuma , casorati determinant solutions for the discrete painlev iii equation , j. math . phys . * 36 * ( 1995 ) 41624174 . k. kajiwara , y. ohta , j. satsuma , b. grammaticos and a. ramani , casorati determinant solutions for the discrete painlev - ii equation , j. phys . * 27 * ( 1994 ) 915922 . k. kajiwara and j. satsuma , @xmath0-difference version of the two - dimensional toda lattice equation , j. phys . japan * 60 * ( 1991 ) 39863989 . r. koekoek , p.a . lesky and r.f . swarttouw , hypergeometric orthogonal polynomials and their @xmath0-analogues , springer monographs in mathematics . springer - verlag , berlin , 2010 . kruskal , k.m . tamizhmani , b. grammaticos and a. ramani , asymmetric discrete painlev equations , regul . chaotic dyn . * 5 * ( 2000 ) 273280 . t. masuda , classical transcendental solutions of the painlev equations and their degeneration , tohoku math . j. * 56 * ( 2004 ) 467490 . mcintosh , some asymptotic formulae for @xmath0-shifted factorials , ramanujan j. * 3 * ( 1999 ) 205214 . m. murata , h. sakai and j. yoneda , riccati solutions of discrete painlev equations with weyl group symmetry of type @xmath372 , j. math . * 44 * ( 2003 ) 13961414 . n. nakazono , hypergeometric @xmath373 functions of the @xmath0-painlev systems of type @xmath17 , _ sigma symmetry integrability geom . methods appl . _ * 6 * ( 2010 ) . n. nakazono and s. nishioka , solutions to a @xmath0-analog of the painlev iii equation of type @xmath374 , funkcial . ( in press ) . s. nakao , k. kajiwara and d. takahashi , on the multiplicative dp@xmath375 and its ultradiscretization , kyushu university ouyorikigakukenkyushukai houkoku , 9me - s2 , evolutionary advancement in theories of soliton ( 1998 ) 125130 ( in japanese ) . m. noumi , painlev equations through symmetry ( american mathematical society , providence , 2004 ) . y. ohta , a. ramani and b. grammaticos , an affine weyl group approach to the eight - parameter discrete painlev equation . symmetries and integrability of difference equations , j. phys . a : math . * 34 * ( 2001 ) 1052310532 . y. ohta , self - dual structure of the discrete painlev equations , rims kokyuroku * 1098 * ( 1999 ) 130137 ( in japanese ) . k. okamoto , studies on the painlev equations . second and fourth painlev equation , @xmath376 and @xmath377 , math . ann . * 275 * ( 1986 ) 221255 . k. okamoto , studies on the painlev equations . fifth painlev equation @xmath378 , japan . j. math . * 13 * ( 1987 ) 4776 . v. periwal and d. shevitz , unitary - matrix models as exactly solvable string theories , phys . * 64 * ( 1990 ) 13261329 . t. prellberg , uniform @xmath0-series asymptotics for staircase polygons , j. phys . a * 28 * ( 1995 ) 12891304 . quispel , j.a.g roberts and c.j . thompson , integrable mappings and soliton equations , phys . lett . a * 126 * ( 1988 ) 419421 . quispel , j.a.g roberts and c.j . thompson , integrable mappings and soliton equations ii , physica d * 34 * ( 1989 ) 183192 . a. ramani and b. grammaticos , discrete painlev equations : coalescences , limits and degeneracies , physica a * 228 * ( 1996 ) 150159 . a. ramani , b. grammaticos and j. hietarinta , discrete versions of the painlev equations , phys . rev . lett . * 67 * ( 1991 ) 18291832 . a. ramani , b. grammaticos , t. tamizhmani and k. m. tamizhmani , special function solutions of the discrete painlev equations . advances in difference equations , iii , comput . appl . * 42 * ( 2001 ) 603614 . a. ramani , y . ohta , j. satsuma and b. grammaticos , self - duality and schlesinger chains for the asymmetric d - p@xmath375 and @xmath0-p@xmath379 equations , comm . phys . * 192 * ( 1998 ) 6776 . h. sakai , casorati determinant solutions for the @xmath0-difference sixth painlev equation , nonlinearity * 11 * ( 1998 ) 823833 . h. sakai , rational surfaces associated with affine root systems and geometry of the painlev equations , comm . * 220 * ( 2001 ) 165229 . tamizhmani , b. grammaticos , a.s . carstea and a. ramani , the @xmath0-discrete painlev iv equations and their properties , regul . chaotic dyn . * 9 * ( 2004 ) 1320 . watson , the continuation of functions defined by generalized hypergeometric series , trans . soc * 21 * ( 1910 ) 281299 . r. wong , asymptotic approximations of integrals , ( society for industrial and applied mathematics ( siam ) , philadelphia , pa , 2001 ) .
we consider the symmetric @xmath0-painlev equations derived from the birational representation of affine weyl groups by applying the projective reduction and construct the hypergeometric solutions . moreover , we discuss continuous limits of the symmetric @xmath0-painlev equations to painlev equations together with their hypergeometric solutions . * hypergeometric solutions to the symmetric @xmath0-painlev equations * + + @xmath1 : institute of mathematics for industry , kyushu university , + 744 motooka , fukuoka 819 - 8581 , japan . + [email protected] + @xmath2 : school of mathematics and statistics , the university of sydney , new south wales 2006 , australia . + [email protected] + * 2010 mathematics subject classification : * 33d15 , 33e17 , 34m55 , 39a13 + * keywords and phrases : * @xmath0-painlev equation ; painlev equation ; basic hypergeometric function ; hypergeometric function ; saddle point method ; projective reduction ; continuous limit +
nonperturbative qcd and the description of the qcd phase diagram are topics of intense research . among the different approaches to nonperturbative qcd , the use of effective models has proven to be a powerful tool in studying thermodynamics and the phase diagram . particularly , the nambu jona - lasinio ( njl ) model @xcite and its nonlocal version ( nnjl ) @xcite have been extensively used for studying thermodynamics of the low energy limit of qcd . + in a few recent articles @xcite , it has been shown that some odd thermodynamical behavior may occur when working with nnjl models . negative pressure and oscillating entropy are some of the problems encountered in these cases . this kind of behavior is refered to as _ thermodynamic instabilities_. in this article , we show that the thermodynamic instabilities are produced by the pressence of some poles in the light quark propagator . the quasiparticle interpretation of the poles of the propagator allows us to comment on the physical meaning of these poles and their relation to the thermodynamic instabilities . + reference @xcite also shows that the inclusion of the polyakov loop in nnjl models , produces a softening effect on the thermodynamic instabilities . here , we consider the effect of the polyakov loop on the poles of the light quark propagator , in order to understand the reason behind the softening of thermodynamic intabilities . we start by considering the nonlocal njl model described by the euclidean lagrangian @xmath0,\ ] ] where @xmath1 is a light quark field of bare mass @xmath2 . nonlocality is incorporated through the currents @xmath3 where @xmath4 . the function @xmath5 in ( [ mcurrent ] ) is called the _ regulator _ of the model . a bosonization procedure is usually performed through the introduction of a scalar ( @xmath6 ) and pseudoscalar ( @xmath7 ) field , and the mean field approximation is taken @xcite . the light quark propagator in minkowski space is then @xmath8 where @xmath9 and @xmath10 is the mean field value of the scalar field . we will encounter singularities for this propagator at @xmath11 . in what follows , we will adopt the interpretation @xmath12 with @xmath13 the constituent mass of the quark and @xmath14 its decay width . although different identifications can be made ( e.g. @xmath15 ) , we will restrain to the one in eq . ( [ mid ] ) for simplicity . we can then distinguish three different kinds of poles in the propagator : _ real poles : _ poles with @xmath16 and @xmath17 are real poles that correspond to free ( deconfined ) particle states . _ well defined complex poles : _ poles with @xmath16 and @xmath18 are interepreted as confined quasiparticles with a finite decay width . _ ill defined complex poles : _ poles with @xmath19 or @xmath20 are objects that can not be clearly identified with particle states . + the usual bosonization procedure and the mean field approximations are then performed @xcite . we finally arrive to the gap equation @xmath21 where @xmath22,\label{gfin}\end{aligned}\ ] ] where @xmath23 is the fermi - dirac distribution , @xmath24 is the momentum in euclidean space , i.e. @xmath25 and , as usual , @xmath26 . in the njl model @xmath10 and the chiral condensate are related . @xmath27 can be obtained then from eq . ( [ potential ] ) and is an order parameter for the chiral phase transition . this means that the overall thermodynamical information of the model is enclosed on the behavior of @xmath27 or , equivalently , in the poles of the propagator @xmath28 . therefore , we will regard any unphysical behavior in @xmath27 as a _ thermodynamic instability_. since @xmath27 behaves like the chiral condensate , it should decrease monotonically with temperature so , any rising of @xmath27 with temperature will be considered as a signal of instability . this kind of instability may then produce unphysical behavior like negative pressure and oscillating entropy @xcite . @xmath27 will be the main quantity on which we will concentrate in studying thermodynamic instabilities . let us start by considering the nnjl model with a gaussian regulator @xmath29 in euclidean space . among the allowed values for the parameters of the model , we consider here two sets chosen to illustrate opposite kinds of analytical structure for the poles of the propagator . .both sets of parameters used for the gaussian regulator case . @xmath30 is the mean value of the scalar field at zero temperature . [ cols="^,^,^,^,^",options="header " , ] then , by solving @xmath31 one can find the poles of the propagator , which means @xmath32 plane . the dashed lines are solutions to eq . ( [ i m ] ) and the solid lines solutions to eq . ( [ re ] ) . ] [ fu ] shows the poles of the propagator for both sets of parameters . parameter set a exhibits an ill defined complex pole with a negative squared mass followed by an infinite set of ill defined complex poles . set b , however , exhibits two real poles followed by an infinite set of ill defined complex poles . the infinite number of ill defined complex poles is not very important since , given that they are much more massive than the first poles , their contribution is neglegible . in fact , we will solve the gap equation considering only the first three poles in each case and the result will not differ in any significant way from the complete solution . + as a function of temperature . the solid line corresponds to set a and the dashed line to set b. in both cases the the first three poles of the propagator are being considered . ] [ f3 ] shows the behavior of @xmath27 for both sets of parameters . while set b shows no sign of thermodynamic instability , set a presents a bump near @xmath33 mev . as we stated above , all of the thermodynamic information of the model is enclosed in the poles of the propagator . this means that the odd behavior exhibited by @xmath10 must be related to the analytical structure of the poles of the propagator . if one removes the ill defined complex pole of set a , then the instability can be corrected . this will be shown more clearly in our next example . + let us now consider a fractional lorentzian regulator of the form @xmath34 in euclidean space . this regulator is inspired by lattice data from the light quark propagator @xcite . in order to study the poles of the propagator we need to wick rotate into minkowski space . this means we need to define what we will understand by the multivalued function @xmath35 , by defining @xmath36 and taking @xmath37 . in this manner we keep the multivalued nature of our functions and the poles of the propagator will occur in two riemman sheets fig . [ f8 ] shows the poles of the propagator for the fractional lorentzian regulator . we have four poles : a well defined complex pole @xmath38 and three ill defined complex poles @xmath39 . one would then expect to have a thermodynamic instability in the model due to @xmath39 . however , if we solve the gap equation considering only @xmath38 no such instability should appear . as a function of temperature . the solid line corresponds to the solution obtained from counting all the poles . the dashed line is the solution from counting only @xmath38 and the dotted line is the solution from counting only @xmath40 , @xmath41 and @xmath42 . ] [ f9 ] shows that the instability is present when considering all poles but dissapears once the ill defined complex poles are neglected . furthermore , the instability grows if only ill defined complex poles are taken into account . we can say then that while ill defined complex poles cause instabilities , real and well defined complex pole exhibit the expected behavior from a condensate . + in @xcite the authors show how these instabilities are softened through the inclusion of the polyakov loop . since we have seen how these instabilities are produced , the polyakov loop should then have a stabilizing effect on the poles in order to soften the instability . in order to see that let us include the polyakov loop in the model and work in the polyakov gauge as in @xcite . the light quark propagator becomes @xmath43}\boldsymbol{k}\label{propol},\end{gathered}\ ] ] where @xmath44 is the only component of the gluon fields that survives after we specify our gauge and @xmath45 is a matrix with no singularities whose explicit form is not relevant for us . in eq . ( [ propol ] ) we see that the usual poles of the propagator are still present , but new poles have been included through the second factor in the denominator . since the polyakov loop softens these instabilities , the quasiparticles associated with the new poles should be more stable than the previous ones . this is illustrated in fig . we have discussed the appearance of thermodynamic instabilities in nnjl models with gaussian and fractional lorentzian regulators . we showed how these instabilities are related to the presence of ill defined complex poles in the propagator of the model . the thermodynamic instabilities can then be eliminated by simply selecting a set of parameters such that the propagator has no ill defined complex poles ( e.g. set b for the gaussian regulator in this article ) . this can be considered as a new requirement for the chosen set of parameters . however , it may be the case that such set of parameters can not be found . in this case one can still eliminate the instabilities by selecting only some poles to be considered in the propagator . this can be done in a well defined manner if one is working in the real time formalism . when calculating the real time thermal propagator from the spectral density function ( sdf ) , the integration contour used to compute the sdf can be modified in order to exclude the ill defined complex poles @xcite . + we also show how the softening effect of the polyakov loop , is related to the inclusion of new poles of the propagator together with the polyakov loop . the quasiparticle states asociated with this new poles are more stable than what is found at vanishing polyakov loop . therefore , the thermodynamic instabilities are softened because the contribution from ill defined complex poles to the gap equation is diminished . the authors would like to acknowledge support from fondecyt under grant no . 1130056 . m.l . also acknowledges support from fondecyt under grant no . 1120770 . f.m . acknoledges support from conicyt under grant no . 21110577 . s. benic , d. blaschke , and m. buballa , phys . d * 86 * , 0740002 ( 2012 ) . s. benic , d. blaschke , and g. a. contrera , phys . d * 89 * , 016007 ( 2014 ) . m. buballa , phys . rep . * 407 * , 205 ( 2005 ) . s. p. klevanky , rev . phys . * 64 * , 649 ( 1992 ) . r. bowler , and m. birse , nucl . a582 * , 655 ( 1995 ) . r. s. plant , and m. birse , nucl . * a628 * , 607 ( 1998 ) . d. gomez dumm , and n. n. scoccola , phys . d * 65 * , 074021 ( 2002 ) . s. noguera , and n. n. scoccola , phys . d * 78 * , 114002 ( 2008 ) . m. loewe , f. marquez , and c. villavivencio , phys . d * 88 * , 056004 ( 2013 ) . f. marquez , phys . d * 89 * , 07610 ( 2014 ) . m. loewe , p. morales , and c. villavicencio , phys . d * 83 * , 096005 ( 2011 ) . t. hell , s. roessner , m. crfistoferetti , and w. weise , phys . d * 82 * , 054026 ( 2010 ) .
it has been recently pointed out , that nonlocal nambu jona - lasinio models , may present unphysical thermodynamical behavior like negative pressure and oscillating entropy . here we show how these thermodynamic instabilities can be related to the analytical structure of the poles of the quark propagator in the model . the analysis is carried out for two different regulators and we show , in each case , how the instabilities are related to the pressence of highly unstable poles . we also argue that the softening of these instabilities by the inclusion of the polyakov loop is related to the effect the latter has on the poles of the propagator . nambu jona - lasinio , nonlocal , thermodynamic instabilities
the casimir force @xcite between a pair of parallel , perfectly reflecting plates is remarkable prediction of quantum electrodynamics which has been confirmed by experiment @xcite . it also has some implications for the semiclassical theory of gravity , as the stress tensor of the casimir energy violates the weak energy condition . simply from casimir s result for the force per unit area , one can construct the entire stress tensor , using conservation , tracelessness and symmetry arguments @xcite . the result is @xmath0 here the plates are separated by a distance @xmath1 in the @xmath2-direction , and units where @xmath3 are used . here @xmath4 is understood to be a renormalized expectation value of the quantum stress tensor operator . because the local energy density is negative , the weak energy condition is violated . in addition , the null energy condition , @xmath5 for all null vectors @xmath6 is violated as well , except for the case where @xmath6 is parallel to the plates , in which case @xmath7 . the null energy condition is the condition for gravity to locally focus a bundle of null rays . the average null energy condition , @xmath8 along a complete null geodesic , is more difficult to violate @xcite . null rays which are not parallel to the plates eventually intersect and pass through the plates . the integral in eq . ( [ eq : anec ] ) then gets a contribution from the matter composing the plates . the extent to which quantum fields could violate the average null energy condition is of interest in several aspects of gravity theory . its violation , for example , is essential to construct traversable wormholes @xcite . some authors @xcite have suggested that when the assumption of perfect conductivity is removed , the negative energy density might disappear . however , in a previous paper @xcite , we calculated the energy density between two half - spaces filled with dispersive material , and showed that the energy density in the center can still be negative . however , the energy density is no longer constant in the region between the interfaces , and diverges positively at the boundaries . this divergence , which arises despite a dispersive dielectric function which approaches one at high frequency , can be attributed to the assumption of a sharp boundary between the dielectric and vacuum regions . it may come as a surprize that the energy density is finite between perfectly reflecting plates , but diverges near plates of finite reflectivity . however , in the perfectly reflecting case there is a cancellation between two divergent terms . both the mean squared electric field , @xmath9 , and mean squared magnetic field , @xmath10 , diverge but the energy density is finite . when the plates have finite reflectivity , both @xmath9 and @xmath10 diverge less rapidly than in the perfectly reflecting case , but the cancellation is upset , so that the energy density also diverges near the plates . there are more complicated geometries where similar cancellation can occur . in the interior of a wedge with perfecty reflecting walls , the energy density diverges near the corner , but is finite if one approaches either wall away from the corner @xcite . again there must be a cancellation which would be upset if the wall had finite reflectivity . in the present paper , we will extend the results of ref . @xcite to study the pressure components of the stress tensor and the effects of finite temperature . as before , our calculations are based on a formalism developed by schwinger , deraad , and milton@xcite . lorentz - heaviside units with @xmath3 will be used . we consider the electromagnetic field stress tensor in the vacuum region of width @xmath1 between two dielectric half - spaces whose dielectric function is described by the plasma model : @xmath11 where @xmath12 is the plasma frequency . the finite mean squared electric field in the vacuum region , at zero temperature , can be expressed as an integral over imaginary frequency @xmath13 @xcite @xmath14 e^{-\kappa a}\cosh\left [ \kappa(2z - a)\right ] \right\}\ , , \end{gathered}\ ] ] where the reflection coefficients for s and p polarizations , respectively , are given by @xmath15 the quantities @xmath16 and @xmath17 are defined as @xmath18 , and @xmath19 . the expression for the mean squared magnetic field @xmath20 is obtained from that for @xmath21 by interchanging the coefficients @xmath22 and @xmath23 . the energy density in the vacuum region , @xmath24 is then @xmath25 \right\ } \label{eq : u } .\end{aligned}\ ] ] as discussed in @xcite , @xmath26 is position dependent : it has a minimum at the center of the vacuum region and diverges at the interfaces . the overall sign of @xmath26 at its minimum depends on the choice of @xmath1 and @xmath12 . as the product @xmath27 grows , @xmath26 at the midpoint decreases , becoming negative for @xmath28 ( see figure 2 , dash - dot line ) . it is of interest to examine the effects due to finite temperature upon the energy density and see when its sign can still be negative when the temperature is not zero . for this purpose , we write eq . ( [ eq : u ] ) as a fourier series instead of an integral on @xmath13 @xcite : @xmath29 \right\ } \ , , \end{gathered}\ ] ] where @xmath30 . the prime on the sum is a reminder to count the @xmath31 term with half weight , and @xmath32 . this expression ( which vanishes as @xmath33 ) represents the casimir contribution to the energy density , or the difference between the energy density at finite temperature with the dielectric walls present and not . it does not include the energy density of a thermal bath without the walls present . to get the latter energy density , we can start with the full expression for @xmath26 , which includes the empty space vacuum divergent term @xcite . at zero temperature , this term can be written as an integral over real frequencies : @xmath34 where @xmath16 is defined as @xmath35 . for finite temperatures , this term is modified by inserting a factor @xmath36 $ ] to account for the thermal energy . this factor reflects the fact that at zero temperature , each mode has an energy of @xmath37 ; at finite temperature , there is an additional thermal energy of @xmath38 . after removing the divergent term , the result is the familiar result for the energy density of blackbody radiation : @xmath39 the energy density in the vacuum region at finite temperature is then@xmath40 [ ptb ] ub5200.eps figure [ fig : ub ] shows two graphs representing energy density at zero and finite temperature corresponding to @xmath41 . as expected , the local energy density increases with temperature , and negative energy density is only possible if the temperature is sufficiently low . on the other hand , the zero temperature results are a good approximation so long as @xmath42 . for systems at room temperature , this increase in the energy density is still very small when the separations between the walls are of the order of few micrometers . more generally , one can ignore thermal effects at distances small compared to @xmath43 . in this case , it is still possible to achieve negative energy density in the central region . at room temperature , for example , @xmath44 . thus , @xmath41 corresponds to @xmath45 , which is in the range of separations for which casimir force experiments have been performed . [ ptb ] uvsobeta2half5.eps the energy density at the center of the vacuum region is shown in fig . [ fig : ucenter ] as a function of @xmath27 for various temperatures . as the temperature increases , the region of negative energy density shrinks , and when @xmath46 , the energy density is positive everywhere . first consider the longitudinal pressure , @xmath47 . the conservation law , @xmath48 , with @xmath49 , and the fact that @xmath50 is diagonal , implies that @xmath51 . thus @xmath52 is constant . from the relation @xmath53 , we find that , at zero temperature , @xmath54 [ ptb ] tzzvsop.eps the plot of this function is shown in fig . [ fig : tzzcenter ] as a function of @xmath27 at any value of z in the vacuum region . the horizontal line corresponds to the perfectly reflecting wall , @xmath55 . note that @xmath56 is the force per unit area on one half - space due to the other , and agrees with the result of the lifshitz theory @xcite . the magnitude of the force is maximum in the perfectly conducting limit . the transverse pressure , @xmath57 is a nontrivial function of @xmath2 , and is given at zero temperature by @xmath58 \right\ } \ , , \label{eq : px}\end{aligned}\ ] ] [ ptb ] txxvsop.eps the plot of this function is shown in fig . [ fig : txxcenter ] as a function of @xmath27 at the center of the vacuum region . the horizontal line corresponds to the perfectly reflecting wall . if we compare the position dependent term in @xmath26 , eq . ( [ eq : u ] ) , with that in @xmath59 , eq.([eq : px ] ) , we see that they differ by a factor of two . these terms are dominant near a wall , so the asymptotic form of @xmath59 near the boundary @xmath60 is one half the corresponding expression for @xmath26 in this limit , which was found in ref . @xcite : @xmath61 again , this singular behavior arises despite the inclusion of dispersion in our treatment , and can be viewed as due to the assumption of a sharp boundary at @xmath62 . a crucial point is that the reflection coefficients vanish as @xmath63 as @xmath64 . pfenning @xcite has studied scalar models in which these coefficients vanish more rapidly at high frequency , and obtained a finite stress tensor at the boundary . now we turn to the null energy condition for rays travelling parallel to the walls . combining eqs . ( [ eq : u ] ) * * ( * * @xmath65 * ) * and ( [ eq : px ] ) , and changing variables using @xmath66 and @xmath67 , we have @xmath68 e^{-ua}\cosh\left [ u\left ( 2z - a\right ) \right ] \right\ } \ , . \label{null}\end{gathered}\ ] ] we see that in case of @xmath69 and @xmath70 , or perfectly reflecting walls , @xmath71 , so the null energy condition is marginally satisfied . we now wish to show that in all other cases , @xmath72 . first consider the terms proportional to @xmath73 . of these , the term proportional to @xmath74 vanishes because @xmath22 is independent of @xmath75 and @xmath76 . next consider the term proportional to @xmath73 and to @xmath77 . its contribution to the integral on @xmath75 can be written , using partial integration , as @xmath78 where in terms of the new coordinates , @xmath79 its derivative , @xmath80^{2}}\,,\ ] ] is negative . thus , @xmath81 now consider the term in eq . ( [ null ] ) proportional to @xmath82 . we can write @xmath83 from which we see that @xmath84 . we can also show that @xmath85 for @xmath86 , from which it follows that @xmath87 . this implies that the @xmath82 term in eq . ( [ null ] ) is non - negative . thus @xmath88 , and the null energy condition is satisfied by a finite margin , except for the limiting case of a perfect conductor . this implies that the gravitational effect on light rays moving parallel to the plates is to cause focusing . even though the energy density can be negative , its effect is more than cancelled by the positive pressure . now we inspect the effect of the finite temperature on the null energy condition . we expect that finite temperature would make the null energy condition satisfied by a wider margin . so , we write eq . ( [ null ] ) as fourier series as above in eq . ( [ uct ] ) : @xmath89 \right\ } + \frac{4\pi^{2}}{45\beta^{4}}.\end{gathered}\ ] ] evaluated at @xmath90 , where it has a minimum , this becomes : @xmath91 a change of variables @xmath92 has been made . a plot of eq . ( [ nulltcenter ] ) as a function of @xmath27 is shown in fig . [ nullb4 ] . the quantity @xmath93 is always positive and increases with increasing temperature , so the null energy condition is always satisfied for transverse rays . [ ptb ] nullb4.eps now we examine the null energy condition for light rays perpendicular to the walls . adding eqs . ( [ eq : tzz ] ) and ( [ eq : u ] ) yields , at zero temperature @xmath94 e^{-ua}\cosh\left [ u\left ( 2z - a\right ) \right ] \right\ } .\end{gathered}\ ] ] [ ptb ] nullz3.eps a plot of eq . ( [ utzz ] ) as a function of position is shown in fig . [ utzz3 ] for three values of @xmath27 . the bottom solid line corresponds to the case of perfectly reflecting walls , @xmath95 , for which the null energy condition in the @xmath2 direction is violated . the figure shows that in the case of dielectric walls , the null energy condition can be violated locally in this direction , but only over a restricted interval in the internal region . the average null energy condition integral , eq . ( [ eq : anec ] ) , will acquire a net positive contribution even before a null ray reaches a boundary . finite temperature will further restrict the region where the null energy condition can be violated . in this paper , we have discussed the effects of finite reflectivity of the walls and of finite temperature on the casimir energy density and pressures . in particular , we have been interested in when the weak energy condition and the null energy condition can be violated . we find that the weak energy condition and the null energy condition for rays not parallel to the plates can still be violated , but with more difficulty than in the case of perfectly reflecting plates . furthermore , these violations are now confined to a localized central region finitely removed from the boundaries . these regions decrease in size as the temperature increases , and eventually vanish for @xmath96 . the energy density and transverse pressure diverge positively as the boundaries are approached , further limiting the region of possible energy condition violation . the null energy condition for rays parallel to the boundary , which is marginally satisfied for the perfectly reflecting case at zero temperature , is satisfied by a finite margin with either finite reflectivity or finite temperature .
the stress tensor for the quantized electromagnetic field is calculated in the region between a pair of dispersive , dielectric half - spaces . this generalizes the stress tensor for the casimir energy to the case where the boundaries have finite reflectivity . we also include the effects of finite temperature . this allows us to discuss the circumstances under which the weak energy condition and the null energy condition can be violated in the presence of finite reflectivity and finite temperature .
materials composed of hard cohesionless grains , such as dry sand , exhibit many remarkable properties ranging from cluster formation in gaseous phases,@xcite to unexpected flows and jets in fluid - like phases,@xcite and to complex organization of inhomogeneous stresses in dense fluid or solid phases.@xcite in this paper we focus on one feature of the solid phase that has received much attention : the distribution @xmath0 of contact forces between grains in a system that is supporting a macroscopic compression or shear force . in a number of experiments and numerical simulations involving non - cohesive grains , @xmath0 appears to decay approximately exponentially at large @xmath1,@xcite which runs counter to the nave expectation of a gaussian distribution about some average @xmath1 . several theoretical analyses of model systems have indicated possible explanations for an exponential tail,@xcite including an analytic calculation for a special case of isostatic packings of frictionless disks . @xcite still , a fundamental understanding of the phenomenon has not been achieved . to motivate the problem considered in the present work , we recall that for generic packings of frictional and/or nonspherical hard particles geometrical constraints permit the formation of more contacts than would be required for supporting imposed stresses . that is , the contact network can contain enough contacts that the stress balance conditions do not determine a unique configuration of the intergrain forces . ( see , for example , ref . @xcite . ) in such cases , the determination of @xmath0 must involve some sort of average over the ensemble of possible force configurations . edwards has suggested that the appropriate measure in configuration space for this ensemble is a flat one ; i.e. , that all possible force configurations should be considered equally weighted.@xcite in general , such an ensemble should include averages over different contact network geometries as well as different stress states on a given network.@xcite in this paper we investigate the question of whether edwards hypothesis leads to exponential tails in @xmath0 for a system of non - cohesive grains in which the contact network forms a triangular lattice . we first study the case of hydrostatic compression , where our results confirm those of snoeijer et al,@xcite though we employ different boundary conditions and different numerical methods . we then examine the effects of diluting the lattice to the rigidity percolation threshold . the transition appears to be first order , with a finite jump in the number of available configurations at threshold . @xmath0 becomes substantially broader than in the full lattice case , but still decays faster than exponentially on the lattice sizes within our numerical reach . finally , we consider the effects of anisotropic imposed stresses . we show that strong anisotropy must produce an exponential tail in large systems and present numerical results showing the approach to this limit for varying degrees of anisotropy . because the difference between the decay produced by our model and a true exponential decay is rather subtle , it is difficult to give firm interpretations of most of the experimental and numerical results . observation of two or three decades of exponential decay could still be consistent with the slowly developing deviation we observe in isotropic systems . on the other hand , a given experiment may involve sufficiently strong anisotropies that we would expect something even closer to a true exponential . the recent results of majmudar and behringer @xcite clearly illustrate that anisotropy due to shearing produces distributions much closer to exponential than those observed under isotropic compression , but we urge caution in making direct comparisons between the model results and experiments on disordered , deformable grains . in the model described below , all forces are directed along the line of contact between two grains and tensile forces are not allowed . this may be thought of as corresponding to the case of frictionless , non - cohesive , circular disks , though a generic set of perfectly circular disks could not form as many contacts as are present in the model . periodic boundary conditions are imposed , so there is no distinction between bulk and boundary contacts . by treating the force network without regard to any distortions of the lattice , we are considering systems of grains with elastic moduli large enough that the boundary forces cause negligible strain ; i.e. , we consider either very hard grains or very weak boundary forces . in all cases , we fix the contact network and study the ensemble of force configurations on that network . we note that fluctuations are large . @xmath0 in any given force configuration may look quite different from the @xmath0 obtained by averaging over configurations , even on the largest lattices we have studied . for the smallest nontrivial lattice , consisting of nine grains , we calculate @xmath0 both analytically and numerically , and find good agreement . for larger lattices , we employ numerical sampling methods involving monte carlo moves that maintain the force balance constraints at all times ( in contrast to simulated annealing methods @xcite ) . we study the distribution of bond strengths on an @xmath2 triangular lattice corresponding to the contacts in a hexagonal packing of monodisperse circular grains . we consider a system in the shape of a rhombus , having @xmath3 bonds or edges in each lattice direction , and subject to periodic boundary conditions , as illustrated in fig . [ fig : lattice ] . each edge carries a scalar variable @xmath1 specifying the magnitude of the force transmitted across a contact . in order to fix the three components of the macroscopic stress tensor @xmath4 we specify the total compressive forces @xmath5 , @xmath6 , and @xmath7 supported along each of the three lattice directions . triangular lattice ( with @xmath8 ) . under periodic boundary conditions light edges on the left are identified with dangling edges on the right and light edges on top are identified with dangling edges on the bottom . the macroscopic stress state is fixed by imposing compressive forces @xmath5 , @xmath6 , and @xmath7 along each lattice direction . ] there are @xmath9 variables in the system . for each grain , the vector force balance condition imposes @xmath10 constraints . these constraints are not all independent , however . the periodic boundary conditions guarantee that the sum of all the single - grain constraint equations is trivially zero , leaving @xmath11 independent constraints . fixing @xmath5 , @xmath6 , and @xmath7 imposes 3 additional constraints , for a total of @xmath12 constraints . this leaves @xmath13 degrees of freedom in the force configuration . in addition to force balance equations , there are inequality constraints associated with the fact that the material is non - cohesive . no force @xmath1 is allowed to be negative . it is this condition that introduces nonlinearity in the system . for any two force balanced , tension free configurations on the lattice , any weighted average with positive weights will also be force balanced and tension free . sums with a negative weighting of a configuration , however , will not always be allowed , as they may contain negative forces on some edges . we represent the system configuration by a vector of forces @xmath14 , @xmath15 . the above constraints create a space of possible configurations that fills a finite , convex @xmath13-dimensional volume , with boundaries determined by the inequalities @xmath16 . our interpretation of the edwards hypothesis is to assign a uniform probability ( lebesgue ) measure @xcite on the space of allowed configurations . in principle , @xmath0 can be calculated in the following way . fix one edge to carry force @xmath1 ; this restricts the system to a ( @xmath17)-dimensional subset of the ( @xmath13)-dimensional allowed volume @xmath18 in configuration space . @xmath0 is simply the ratio of the ( @xmath17)-dimensional `` area '' to the ( @xmath13)-dimensional total volume . more formally , let @xmath19 be the set of 6 edges touching node @xmath20 ; let @xmath21 , @xmath22 , be a unit vector along edge @xmath23 in @xmath19 pointing towards node @xmath20 ; and let @xmath24 , @xmath25 , be the set of @xmath3 edges along lattice direction @xmath26 . then we have @xmath27 \left[\prod_{k=1}^3 \delta \left(n f_k-\sum_{f_l \in l_k}{f_l}\right)\right ] \label{eqn : p}\ ] ] where edge @xmath28 has been fixed to have value @xmath1 . the integral is taken over all @xmath1 s . the first delta function ensures that edge @xmath28 carries force @xmath1 . the next ensures force balance at each vertex , and the last enforces the boundary conditions . the integral represents the volume of a @xmath29-dimensional slice of a polytope , therefore @xmath0 on an @xmath30 lattice is a piecewise polynomial of order @xmath17 . the average force on the edges in @xmath24 is @xmath31 . the last factor in eqn . ( [ eqn : p ] ) requires comment . for definiteness , consider @xmath32 . let us divide the set of edges in @xmath33 into @xmath34 layers , each layer containing the @xmath34 edges that intersect a line along one of the other edge directions . the following argument shows that the sum of the @xmath35 on each layer must be equal to @xmath6 . [ fig : slab ] shows a shaded slab of the system . the total force on this slab must be zero , which implies that the vector sum of the forces on the edges indicated by thick lines on one side of the slab must be equal to the vector sum of the forces on the other side . since the vector sum has a unique decomposition into contributions from the @xmath33 edges and the @xmath36 edges , the sum of the @xmath35 s in each direction must independently be the same on both sides of the slab . this argument is independent of the thickness of the slab , so in the @xmath33 direction each of the @xmath34 layers must have @xmath35 s that sum to @xmath6 , and similarly for @xmath37 and @xmath36 . must have a sum of forces equal to @xmath38 . ] for the @xmath39 case , the integral can be evaluated analytically . here we state the isotropic result ; the anisotropic result is presented below . the calculation is detailed in appendix a. ( we have not found an analytic expression for @xmath0 for arbitrary lattice size . see ref . @xcite , however , for a treatment of the case in which the sum @xmath40 is fixed , but not the individual @xmath38 . ) in the isotropic case ( @xmath41 ) , we find @xmath42 where @xmath43 is the heaviside step function . this expression is plotted in fig . [ fig : iso3x3 ] on larger lattices we employ numerical methods to measure @xmath0 . this requires generating numerous configurations in the allowed volume of configuration space in a manner consistent with the edwards flat measure . previous work has implemented a simulated annealing algorithm , which generates each configuration by starting from a random point in the space of all possible @xmath14 and relaxing to some point in the lower dimensional subset of interest.@xcite we employ a different technique in which , starting from one force - balanced configuration a point in the allowed subset of stress states new configurations are generated via moves that always produce allowed configurations . by identifying a set of moves that span this compact space , reach any set of positive volume in a finite time , and involve symmetric transition probabilities , we can be sure that at sufficiently long times the space is being sampled with uniform measure.@xcite these moves are described below . for the @xmath30 lattice with @xmath44 , we construct a set of @xmath3 _ wheel moves _ , one centered on each node . the wheel move associated with a given node acts on the nearest neighbor edges ( `` spokes '' ) and next nearest neighbor edges ( `` rim '' ) of the node , reducing one set of @xmath1 s and augmenting the other by the same amount . , chosen to be small enough that no edge becomes negative . force on each edge marked with a thick solid segment is increased and on each edge marked with a thick dashed segment is decreased , or vice versa . ] the move is implemented in the following steps . identify the smallest force on the spokes and call it @xmath45 ; likewise , call the smallest force on the rim @xmath46 . randomly choose a force increment @xmath47 with uniform measure on the interval @xmath48 $ ] . adding @xmath49 to every edge on the spokes and subtracting it from every edge on the rim , which respects force balance on every node touching the wheel , constitutes a wheel move . in the space of linear combinations of wheel moves , there is one obvious null direction ; i.e. , a linear combination of wheel moves that results in no change to any edge of the lattice . we prove in appendix b that the number of linearly independent wheel moves is exactly @xmath13 , which is the number required to span the space of force configurations . fig . [ fig : stuck ] illustrates a geometric subtlety that must be taken into account when attempting to explore the allowed stress states using wheel moves ( or any other basis set of moves ) . since the given pair of basis vectors span the two - dimensional convex space shown , it is clearly possible to navigate from any interior point to any other by making discrete steps along the basis vectors . it may not be possible , however , to make a move in either basis direction if the initial configuration lies exactly on a corner . in higher dimensions it is likewise possible to be stuck on a corner or a boundary of higher dimension . it is therefore important to find initial configurations that lie in the interior of the space , rather than on a boundary , for the purposes of our monte carlo sampling technique . an algorithm such as the one described above , beginning at an interior point , can come arbitrarily close to all boundaries but will never reach them . regions near corners are visited infrequently but for long times in such a way that , for sufficiently long runtime , the space is sampled with uniform measure . @xcite fig . [ fig : iso3x3 ] shows the agreement between the exact calculation and simulation on the @xmath50 lattice . the forces are normalized by choosing @xmath51 such that the average force @xmath52 is unity . the peak near @xmath52 is typical for larger lattices as well , but finite size effects are clearly evident in the tail , since @xmath0 must go to zero at @xmath53 ( here @xmath54 ) . triangular lattice under isotropic stress . the curve gives the force distribution , points represent numerical simulations of the same quantity . @xmath0 vanishes at @xmath51 , the force imposed along each lattice direction . here @xmath54 , fixing @xmath55 . ] a typical configuration for the @xmath56 lattice is shown in fig . [ fig : isolatt ] . lattice . the force on an edge is redundantly mapped to color and width ( stronger forces are darker and thicker ) . ] @xmath0 s for the cases @xmath57 , @xmath58 , @xmath59 , and @xmath60 are shown in fig . [ fig : nxn ] . lattices under isotropic stress for @xmath57,@xmath58,@xmath59 , and @xmath60 . for all cases the decay is faster than exponential . the @xmath61 and @xmath62 cases are nearly identical , indicating convergence to a universal curve for large systems . the straight line is drawn as a guide to the eye . for each curve , @xmath55 . ] the peak near @xmath52 is again apparent , and for small @xmath1 the curves appear to coincide , though small differences exist that are masked by the logarithmic scale . for @xmath63 the curves separate , with the @xmath57 distribution decaying most rapidly . the four distributions in the figure all decay faster than exponentially , and they broaden slightly with increasing @xmath34 . to gain confidence that the curves are converging to a large system limit , we measured correlations between forces as a function of the distance between edges . in directions both longitudinal and transverse to a particular edge , as shown in fig . [ fig : correlation ] , we see that on a @xmath64 lattice correlations have decayed to the @xmath65 level at a distance of @xmath58 lattice constants . on larger lattices we see exponential decay with a decay length of approximately @xmath66 lattice constants , although much more data on larger lattices would be necessary to measure this precisely . on a @xmath64 lattice . the correlation is computed for edges that have the same orientation . the longitudinal correlation refers to displacements by @xmath67 lattice constants along the direction of the edge . the transverse correlation refers to displacements perpendicular to the edge . the plots above are averaged over multiple reference edges . the transverse correlations are in fact negative , but we take their absolute value to facilitate plotting on a log scale . ] on a @xmath64 lattice . the correlation is computed for edges that have the same orientation . the longitudinal correlation refers to displacements by @xmath67 lattice constants along the direction of the edge . the transverse correlation refers to displacements perpendicular to the edge . the plots above are averaged over multiple reference edges . the transverse correlations are in fact negative , but we take their absolute value to facilitate plotting on a log scale . ] we therefore expect to see little difference in @xmath0 for lattices larger than @xmath62 . supporting this expectation , there is very little difference between the @xmath61 and @xmath62 distributions for @xmath1 up to the largest values we have measured , which covers five decades of @xmath68 . though it is conceivable that the asymptotic form of @xmath0 at large @xmath1 in our model is exponential , the domain displayed in fig . [ fig : nxn ] is the relevant one for comparison with experiments , and it clearly shows a decay that is faster than exponential . real disks are not perfectly monodisperse , so in any hexagonal packing of hard disks some of the bonds on the triangular lattice will not actually be present . for a given stress state , a diluted lattice has fewer bonds to carry the same force . those remaining will , on average , carry more force , resulting in a shift in @xmath0 toward larger forces , which may involve a broadening of the tail . as the number of edges removed from the lattice increases , the force configurations tend to become less homogeneous , with strong forces concentrated in chain - like structures . lattice at rigidity percolation . deleted edges are covered with disks . the force on an edge is redundantly mapped to color and width ( stronger forces are darker and thicker ) . ] as more edges are deleted there comes a point where the lattice can no longer support the imposed stress ; this is the rigidity - percolation transition . [ fig : perlatt ] shows a typical force configuration for a randomly diluted lattice . we study randomly diluted lattices at the transition to see whether the broadening takes the form of an exponential decay . the random dilution process merits further remark . for a random process in which each bond is removed with probability @xmath69 , the infinite triangular lattice under isotropic compression can not support stress for any @xmath70 . @xcite thus in an @xmath30 simulation the fraction ( not the number @xmath71 ) of deleted edges at the rigidity percolation threshold goes to zero as @xmath72 . as an arbitrarily large real hexagonal packing does support compression , this means that the process of bond breaking in the real packing does not happen randomly but rather in a correlated way . nevertheless , our simulations employ random dilution on finite lattices . we also neglect for present purposes the possible buckling instability of a force chain . disallowing configurations that might buckle could only result in ensembles with fewer dilutions and therefore would not change the conclusions described below . we construct lattices at the rigidity percolation threshold in the following way . given the imposed macroscopic stress and beginning with an empty lattice , edges are selected at random to be added to the lattice . after each addition , we check to see if the lattice is capable of supporting the imposed stress using the simplex algorithm.@xcite we take the lattice to be at threshold when it first supports the imposed stress , at which point we know that there exists at least one edge such that , when removed , the system could no longer support the stress . the edges that have _ not _ been added to the lattice at this stage are called `` deleted edges . '' our construction yields a lattice in which deletions are randomly distributed in space . note that the process will in general yield some edges that are present but not connected to the edge network in a way that allows them to bear any force . we refer to these edges as `` effectively deleted . '' these effectively deleted edges on a lattice supporting compressive forces can be identified in the following way . consider the @xmath73 edges that meet at a given node . an edge becomes effectively deleted if its opposite edge has been deleted ( possibly effectively ) and at least one of the edges making a @xmath74 angle with it has also been deleted ( possibly effectively ) . under these conditions any force on the edge in question can not be balanced by positive forces on the other edges sharing the node . to find all of the effectively deleted edges we examine all nodes repeatedly until no new deletions are found . the wheel moves used for investigation of the undiluted lattice no longer work on the diluted lattice . the wheel moves add or subtract a quantity @xmath49 to each edge they touch , but this is impossible if one of the edges is deleted . therefore each deleted edge renders unusable the four wheel moves to which it belongs . to salvage a set of moves that span the appropriate space , linear combinations of the wheel moves can be formed that leave the deleted edge unaffected . for a single deleted edge there are three linearly independent combinations of the four wheel moves that edge touches which preserve the deleted edge . the deletion has reduced the available degrees of freedom by one . when many edges have been deleted , more complicated linear combinations of wheel moves can be found that preserve the vanishing force on all deleted and effectively deleted edges . we refer to a particular linear combination of wheel moves that remains a viable move on a diluted lattice as a `` multi - wheel move . '' let the number of linearly independent multi - wheel moves be @xmath75 . just as the undiluted lattice had @xmath3 wheel moves and @xmath13 degrees of freedom , the diluted lattice having @xmath75 multi - wheel moves has @xmath76 degrees of freedom . finding a complete set of multi - wheel moves then provides not only a means to perform numerical simulation but also a count of the degrees of freedom in the diluted system . a linearly independent set of @xmath75 multi - wheel moves on the diluted lattice can be constructed as follows . for @xmath77 , form a vector @xmath78 such that the components of @xmath79 specify the effect of the wheel move centered on node @xmath80 on the _ deleted _ edges . that is , @xmath81 if the @xmath82 deleted edge is a spoke of node @xmath80 , @xmath83 if the @xmath82 deleted edge is a rim of node @xmath80 , and @xmath84 otherwise . form the @xmath85 matrix @xmath86 . consider a linear combination of wheel moves with coefficients @xmath87 , @xmath88 . this linear combination has no effect on deleted edges if and only if @xmath89 . thus @xmath90 , where @xmath91 is the null space of @xmath92 . note that we demand the wheel moves respect both real and effective deletions . were we not to specify effective deletions as well , the algorithm would in general produce additional multi - wheel moves which were nonzero on some effectively deleted edges . this is because the multi - wheel moves could also act on a lattice that supports tensile as well as compressive forces . on such a lattice there are fewer effective deletions ; in particular , a node could be force - balanced having only three edges all on the same side of a line through the node . to model a noncohesive material , then , we must include these additional effective deletions as well . we use the following technique to investigate the degrees of freedom in lattices at threshold . edges are added to a lattice one at a time until the rigidity percolation threshold is reached , the set of wheel moves on that threshold lattice are constructed to determine the dimension of the volume of allowed configurations in state space @xmath76 , and the process is repeated for a number of threshold lattices . we refer to the system as having @xmath76 degrees of freedom , as this is the number of multi - wheel moves it possesses . these lattices have differing numbers of deleted edges . for lattice with the _ same _ number of deleted edges @xmath71 we calculate the average ratio of the number of degrees of freedom of the dilute system to the number in an undiluted one : @xmath93 . figure [ fig : transition ] shows @xmath94 plotted against the fraction of of deleted edges @xmath95 . a discrete jump in @xmath96 emerges as @xmath69 is decreased for systems larger than @xmath97 , indicating that the transition is first order . this suggests that @xmath0 near the transition point is unlikely to show qualitatively different behavior from that of the undiluted lattice , though a quantitative broadening of the force distribution is expected . vs. the ratio of deleted edges in the threshold lattice to the total number of edges @xmath98 . for each plot a number of lattices at the rigidity - percolation threshold were generated and the degrees of freedom and deleted edges counted . for large enough lattices , the system has a large number of degrees of freedom available as soon as rigidity - percolation is reached . note that the observed interval of @xmath69 values shifts as the lattice size is increased . ] we measure @xmath0 averaged over a number of threshold @xmath99 lattices . the resulting distribution is indeed much broader than the undiluted case , but there is still curvature in the distribution on a linear - log plot . numerical studies in this regime are hampered by two factors . first , each multi - wheel move requires the examination of many edges to ensure that no @xmath35 becomes negative . second , the maximum size of a multi - wheel move is typically quite small because the move touches several edges that carry small forces . the volume that we are attempting to sample with uniform measure is still convex but contains many corners that are not easily accessed . the data obtained for a given lattice may exhibit a bias depending on the initial configuration unless the number of moves considered is extremely large . we have gathered data using two procedures . in one case we construct a large number of threshold lattices , find a single initial configuration for each and collect force data from @xmath100 wheel moves , then average all of the data together to determine @xmath0 averaged over lattices . to find initial configurations we use the simplex method several times with different coefficients and average the results , thereby avoiding configurations that lie on the boundary of the allowed volume . we can not be sure , however , that this method is free of systematic bias . in the second case , we consider a single lattice at threshold ( with a typical number of degrees of freedom ) , find 25 different initial configurations and run @xmath101 multi - wheel moves from each one , averaging the force data to obtain the @xmath0 associated with that particular lattice . the results are shown in fig . [ fig : perc ] . as may be expected given the first - order nature of the rigidity percolation transition , neither procedure produces an exponential tail in @xmath68 at large @xmath1 . we conclude that random dilution to the rigidity percolation threshold does not produce exponential tails for triangular lattices . for @xmath102 triangular lattices diluted to the rigidity - percolation threshold . the plus signs represent an average over @xmath101 moves on a single randomly diluted lattice . the open circles represent an average over 75 lattices , each run for @xmath100 moves . the straight line is a guide to the eye . ] we now consider the _ undiluted _ lattice with anisotropic stresses imposed by choosing @xmath5 different from @xmath6 and @xmath7 . for example , one lattice direction may be subject to stronger compression than the others , creating qualitatively distinct force distributions in the strong and weak directions . we will show that in strongly anisotropic systems the strong direction contributes an exponential decay to the tail of @xmath0 . we consider @xmath103 , @xmath104 , and parameterize the anisotropy by @xmath105 , so that @xmath106 corresponds to isotropic stress . in terms of @xmath52 , @xmath107 . the average force in the weak direction @xmath108 , while that in the strong direction is @xmath109 . similarly , forces in the weak direction can not exceed @xmath51 , and those in the strong direction can not exceed @xmath110 . forces in the strong direction will populate the tail of @xmath0 . in the limit @xmath111 the system becomes much simpler : only the strong direction carries force , doing so via @xmath34 chains of @xmath34 edges , each edge in a given chain carrying the same force as shown in fig . [ fig : anisolimit ] . the forces on the @xmath34 chains must sum to @xmath112 . limit of the @xmath30 triangular lattice . dashed edges represent attachments under periodic boundary conditions . the macroscopic stress state is fixed by imposing compressive force @xmath112 along the strong lattice direction , the only one carrying force . force is transmitted along @xmath34 chains , each composed of @xmath34 edges , in any way such that a sum over chains yields @xmath112 and all forces remain compressive . ] simple dimensional analysis of the scaling of the allowed volume when one chain is fixed at force @xmath1 yields @xmath113 we see that in this limit anisotropy induces an exponential tail in @xmath0 . we next investigate the extent to which the limiting behavior is reflected at finite anisotropies . to gain some intuition for the approach to the anisotropic limit we return to the @xmath39 case , which can be solved analytically . we let @xmath114 and @xmath115 denote the separate distributions of forces in the strong and weak directions , respectively . @xmath115 is identical to @xmath0 in eqn . [ eqn : isodist ] . in the strong direction , however , the situation is more complicated . a complete expression is given in appendix c. for the @xmath50 case , the fully anisotropic limit @xmath116 has @xmath117 , @xmath118 , a piecewise linear function . triangular lattice under anisotropic stress . the force distribution @xmath114 for forces in the strong direction is plotted for increasing anisotropy . each distribution is scaled to @xmath119 . the limiting form is a piecewise linear function of @xmath1 . @xmath120 for finite anisotropy has a linear region in the middle ; this region grows in width and approaches the limiting slope as anisotropy is increased . ] for @xmath121 , @xmath0 is given by @xmath122 ( from appendix c ) between @xmath51 and @xmath112 . @xmath123 is a linear function , and the interval from @xmath51 to @xmath112 grows with @xmath124 . as @xmath124 increases further the width of this region grows and the slope approaches the limiting slope of @xmath125 . @xmath114 is plotted for increasing @xmath124 in fig . [ fig:3x3aniso ] . on larger lattices , numerical calculations show similar behavior to the @xmath39 case , but with the limiting linear distribution replaced by the limiting exponential of the large system limit over the range of forces of interest . ( the distribution must be cut off at @xmath126 . ) [ fig : anilatt ] shows a typical anisotropic configuration . anisotropic lattice with @xmath127 . the force on an edge is redundantly mapped to color and width ( stronger forces are darker and thicker ) . the stronger stress is imposed in the horizontal direction . ] fig . [ fig : directions ] shows @xmath0 for anisotropic lattice of size @xmath128 , with the contributions from the weak and strong directions shown separately . the role of the strong direction in broadening the distribution is clear . results are shown for two cases , one in which strong forces are imposed on a single lattice direction and another in which strong forces are imposed on two lattice directions . note the difference in behavior for small @xmath1 . in the latter case there is no peak in @xmath0 for small @xmath1 . for strong , weak , and all lattice directions on a @xmath56 lattice with @xmath129 . the strong direction(s ) is entirely responsible for the tail in @xmath0 . top : two weak @xmath38 s and one strong . bottom : one weak @xmath38 and two strong . on both plots , @xmath55 . ] fig . [ fig : anisoexp ] shows force distributions on a @xmath56 lattice for several values of @xmath124 . for sufficiently large @xmath124 a middle portion of @xmath0 appears to be exponential . as @xmath124 increases , so does the extent of this portion . we hold @xmath130 fixed so that every curve will have the same decay length . on a @xmath56 lattice for increasing anisotropy . portions of the tail for stronger anisotropies are difficult to distinguish from exponential decay , and the limiting behavior is a true exponential ( gray line ) . the dashed line indicates the interval in @xmath1 over which the @xmath131 curve is approximately exponential . @xmath132 for every curve . ] we conclude that @xmath0 on the anisotropic lattice displays an exponential tail in the following sense . for moderate @xmath124 and @xmath34 , a portion of the tail of @xmath0 is nearly linear on a log plot . the decay is not truly exponential because of the finite anisotropy and finite lattice size , but the limiting behavior of @xmath0 as @xmath34 and @xmath124 are increased is a true exponential . the region of nearly exponential decay grows with increasing anisotropy , extending as far as @xmath133 for @xmath131 . we note that the degree of anisotropy represented by a given value of @xmath124 can be compared to the anistropy supported by materials described by an internal friction parameter . the ratio of major and minor principal stresses on our lattice is @xmath134=@xmath135 . by exploiting the relation @xmath136 @xcite , where @xmath69 is the angle of internal friction , we find that @xmath137 corresponds to an internal friction of approximately @xmath138 . the survival of the exponential tail of strong forces in large triangular lattices under anisotropic loading can be traced to the discussion based on fig . [ fig : slab ] concerning the sums of forces along layers of edges in the various directions . that argument shows that the strong forces in one lattice direction can not be redistributed into the other directions due to vector force balance constraints . thus for strongly anisotropic lattices , the strong forces will follow scaling laws close to the limiting exponential form corresponding to no force at all in the weak direction(s ) . note that in the case of two strong directions and one weak , the two strong are effectively independent since strong chains from the two directions can not interact significantly without generating a strong chain in the supposedly weak direction and thereby reducing the degree of anisotropy . we have investigated the distribution @xmath0 of contact forces on a lattice of triangular bonds under the edwards flat measure . an interesting question is whether one expects exponential decay in @xmath0 . the triangular lattice is a simple but nontrivial system for studying this phenomenon . the distribution on the @xmath30 periodic lattice decays faster than exponentially , as reported previously.@xcite diluting the lattice induces significant broadening in @xmath0 , but the decay remains faster than exponential . even at rigidity percolation associated with random bond dilution of the lattice , no qualitative change in the form of @xmath0 is discernible . in particular , we do not see evidence for an exponential tail associated with the transition . consistent with these direct measurements of @xmath0 , we find that the transition is first order , which generally suggests that no qualitative changes should be expected as the percolation threshold is approached . on the other hand , imposing anisotropic stress on the undiluted lattice can induce an exponential tail in @xmath0 . in the limit of an infinite lattice with stress imposed only along one lattice direction , the distribution of contact forces is a pure exponential . numerical simulation shows that evidence of this behavior may still be seen for finite lattice sizes and a finite ratio of the compressive forces in the strong and weak directions . in such a scenario the tail of @xmath0 is not a true exponential , but appears approximately linear on a log plot of @xmath0 for some interval in @xmath1 . for large enough anisotropies , @xmath0 decays three orders of magnitude from its maximum and @xmath139 before deviation from exponential decay becomes obvious . in the triangular lattice model , anisotropy is associated with the appearance of long force chains oriented along the strong direction , a phenomenon that has also been observed in experiments on disordered systems.@xcite extension of the numerical techniques employed above to rigid bars that form a disordered triangulation of the plane would be straightforward . wheel moves , for example , would be defined by taking all bars sharing a given vertex to be the spokes , though it would no longer be true that all bars receive force increments of the same magnitude during a move . the primary difference between the prefect triangular lattice and disordered ones ( or even other crystalline ones ) is that lattices with no sets of collinear bars that span the system can not support arbitrarily strong anisotropic stresses ; i.e. , in most cases there would be a maximal value of @xmath124 . a detailed investigation of the effects of disorder would be of interest . @xmath0 for the @xmath50 isotropic lattice is calculated as follows . there are @xmath140 degrees of freedom ; we take nine basis elements subject to a constraint which is to be imposed by hand . three of these elements are shown in fig . [ fig : basis ] . there are three honeycomb elements , @xmath141 , @xmath142 , three elements @xmath143 , @xmath144 , and three elements @xmath145 . the figure depicts @xmath146 , @xmath147 , and @xmath148 . the @xmath149 elements make no net contribution to the total force in the system . the three @xmath69 elements must sum to @xmath51 ; this is the additional constraint we impose . all elements are isotropic by construction . for the @xmath50 isotropic lattice . solid bars indicate a positive contribution to the force on an edge ; dashed bars indicate a negative contribution . double bars indicate a contribution with twice the weight . the honeycomb in ( a ) is one of three . the elements in ( b ) and ( c ) are each repeated for the other two lattice directions . ] [ eqn : p ] can be rewritten @xmath150 where @xmath151 . the first @xmath152-function fixes a particular edge to support force @xmath1 ; all edges are equivalent by symmetry . the bounds on the @xmath149 s assure that no edge will support tensile force . the presence of @xmath153 breaks the integral into three regions . the regions wherein @xmath154 and @xmath155 are identical ; the region where @xmath156 must be evaluated separately because @xmath146 touches the edge fixed to carry @xmath1 . the evaluated expression is given in eqn . [ eqn : isodist ] . adapting our notation from section iv , for each @xmath77 let @xmath157 be the vector that specifies the effect of the wheel move centered at node @xmath80 on the collection of _ all _ edges . form the @xmath158 matrix @xmath159 . a linear combination of wheel moves with coefficients @xmath87 , @xmath88 has no effect on any edge if and only if @xmath89 . _ lemma _ : @xmath91 is spanned by a single vector such that @xmath160 for all @xmath28 , @xmath80 . _ remark _ : since w is @xmath158 , the system of equations @xmath161 is highly overdetermined . _ proof _ : given any edge , let @xmath28 and @xmath80 be the two vertices at the ends of the edge , and let @xmath26 and @xmath162 be the nearest two vertices on the perpendicular bisector ( see fig . [ fig : appb]a ) . the wheel move with coefficients @xmath163 has no effect on the given node if and only if @xmath164 to interpret eqn . [ eqn : b1 ] geometrically , let @xmath165 be the coordinates of node @xmath28 in fig . [ fig : appb]a , and define @xmath166 by @xmath167 define @xmath168 , @xmath169 , and @xmath170 similarly . then eqn . [ eqn : b1 ] is satisfied if and only if @xmath170 lies in the plane in @xmath171 determined by @xmath172 , @xmath168 , and @xmath169 . suppose the wheel move with coefficients @xmath163 has no effect on any edge : i.e. , @xmath173 . order the nodes as indicated in fig . [ fig : appb]b . for the moment we regard the first three coefficients @xmath174 , @xmath175 , @xmath176 as arbitrary . applying eqn . [ eqn : b1 ] to nodes 4 , 5 , and 6 , we deduce that @xmath177 , @xmath178 , and @xmath179 all lie in the plane of @xmath180 , @xmath181 , and @xmath182 . with further application of eqn . [ eqn : b1 ] we can extend the conclusion to all vertices of the outer triangle , and this may be continued to conclude that all points @xmath172 lie in a single plane . finally , periodicity requires that this plane is horizontal : i.e. , @xmath183 for all @xmath28 and @xmath80 . the calculation of @xmath0 for the anisotropic @xmath39 case proceeds similarly to the isotropic calculation . the anisotropy requires the division of the relevant integral according to whether the force @xmath1 is greater than or less than the difference @xmath112 between the strong @xmath184 and the weak ones . for the case @xmath185 , we find @xmath186 where @xmath18 is a normalization constant and the @xmath120 s are given below . for the case @xmath187
we study the uniformly weighted ensemble of force balanced configurations on a triangular network of nontensile contact forces . for periodic boundary conditions corresponding to isotropic compressive stress , we find that the probability distribution for single - contact forces decays faster than exponentially . this super - exponential decay persists in lattices diluted to the rigidity percolation threshold . on the other hand , for anisotropic imposed stresses , a broader tail emerges in the force distribution , becoming a pure exponential in the limit of infinite lattice size and infinitely strong anisotropy .
electroweak symmetry breaking sectors which are composed of two scalar isospin - doublets predict charged higgs bosons ( @xmath1 ) . their discovery at future colliders will directly indicate such a non - minimal structure of the higgs sector , and its detailed information will give strong hints for the structre of physics beyond the standard model ( sm ) . therefore , their phenomenology has been evaluated especially over the past few years as their discovery potential has become more clear . at cern large hadron collider , charged higgs bosons below the top quark mass will be produced in the top quark decay process . for heavier charged higgs bosons , they will be produced mainly through the subprocesses @xmath2 , @xmath3 and their charge conjugate , and they may be probed , for example , using the decay modes @xmath4 and its charge conjugate especially for large @xmath5 values , in spite of huge qcd backgrounds . on the other hand , at future electron - positron linear colliders ( lc s ) , the dominant production process should be the pair production@xcite @xmath6 . its cross section depends only on the charged higgs mass , and this process provides a hallmark channel through which we can study @xmath1 phenomenology@xcite as long as the charged higgs boson mass ( @xmath7 ) is sufficiently smaller than its kinematical threshold @xmath8 . when @xmath7 is near or above @xmath8 , the pair production is not available . in such a case , the phenomenology of charged higgs bosons may be explored through the kinematically - allowed single production processes . in this talk , we discuss various single @xmath1 production channels to complement the pair production channel . we consider the following processes@xcite : @xmath9 the @xmath10 associated production mode ( [ proc_tau ] ) has been studied in the literature@xcite . the processes ( [ proc_tb ] ) , ( [ proc_lwh ] ) , ( [ proc_bwh ] ) and ( [ proc_awh ] ) , as well as ( [ proc_zwh ] ) with @xmath11 lead to the same final state @xmath12@xcite . the one - loop induced processes @xmath13 have also been studied in refs.@xcite . the one - loop induced @xmath14 and @xmath15 vertices which enter into ( [ proc_enu ] ) have been calculated in ref.@xcite . we evaluate cross sections of the above single production processes at the leading order . the two - higgs - doublet - model ( 2hdm ) parameters are taken assuming the minimal supersymmetric standard model . for the sm parameters we adopted the following numbers : @xmath16 gev , @xmath17 gev , @xmath18 mev , @xmath19 gev , @xmath20 , @xmath21 gev , @xmath22 gev , @xmath23 gev , @xmath24 gev , @xmath25 . the center - of - mass energy of the collider ( @xmath26 ) is assumed to be @xmath27 gev and @xmath28 gev . details of the calculation and the plots are shown in ref.@xcite . the @xmath10 associated production processes ( [ proc_tau ] ) , shown in fig . 1 , are dominated by the pair production as long as it is kinematically allowed ( @xmath29 ) . above this threshold , the cross sections still exceed @xmath30 pb at large @xmath5 values . if @xmath1 is heavy enough to decay into @xmath31 , dominant background contribution comes from top - quark pair production . by peforming @xmath32 and @xmath33 mass reconstruction , the background can be reduced naively by @xmath34 , so that the signal would be visible , although detailed simulations are needed depending on each machine . + + the @xmath35 ( or @xmath36 ) associated production process ( [ proc_tb ] ) has a similar structure to @xmath37 modes , but the @xmath5 dependence is opposite : the rate is larger for smaller @xmath5 values . the cross section is suppressed above the pair production threshold in comparison with that of @xmath37 modes because of the difference in the phase space . figure 2 shows the cross section of the one - loop induced processes @xmath38 ( [ proc_wh ] ) . for small @xmath5 values , the @xmath39 coupling included in each quark - loop diagram is large so that the cross section becomes substantial : the @xmath5 dependence of the cross section is @xmath40 at small @xmath5 and @xmath41 at very large @xmath5 . the cross section remains large beyond the pair production kinematic limit . when the dominant @xmath1 decay mode is @xmath42 , we can deal with the top pair background by reconstructing the final state and eliminating events with two top quarks . this is expected to reduce the top pair background by about @xmath43 , which should enable the observation of the peak at the charged higgs mass in some regions of the parameter space . for the rest of the processes , ( [ proc_enu])-([proc_enah ] ) , there is a variety of structures observed in the cross sections . their rates , however , turned out to be very small for larger @xmath7 than the pair production threshold . this is owing to the reasons such as coupling suppression , cancellation between feynman diagrams , and/or smaller phase space . details of the numerical results are shown in ref.@xcite . as mentioned , the processes ( [ proc_tb ] ) , ( [ proc_zwh ] ) , ( [ proc_lwh ] ) , ( [ proc_bwh ] ) and ( [ proc_awh ] ) lead to the final state @xmath44 . some channels have large cross sections for @xmath29@xcite . the variety of resonance structures imply little interference among these processes . the @xmath45 modes , therefore , complement the pair production in the charged higgs study at lc especially below the threshold of the pair production@xcite . the procedure for the tagging of the final state @xmath44 is outlined in ref.@xcite . in either process , information of the tau@xcite and the top@xcite polarization can be used to help identify the charged higgs boson . we found that the @xmath46 channel and the loop induced @xmath47 channel are the most promising channels for studying charged higgs phenomenology beyond the kinematic limit for pair production . the @xmath47 channel is enhanced at low @xmath5 , whereas the @xmath46 channel is enhanced at large @xmath5 . for the charged higgs bosons whose mass is below the pair production threshold , some of the single production channels which lead to the @xmath44 final state complement the pair production . some of their cross sections become large , but they are rapidly suppressed above the kinematic limit of the pair production . our results motivate further study of the decay modes , the hadronic final states and the backgrounds relevant to these charged higgs boson single production processes . rizzo , _ mod . * a4 * ( 1989 ) 2757 ; a. mendez and a. pomarol , _ nucl . phys . _ * b349 * ( 1991 ) 369 ; m. capdequi peyranre , h.e . haber and p. irulegui , _ * d44 * ( 1991 ) 191 ; s. kanemura , _ phys . rev . _ * d61 * ( 2000 ) 095001 .
we discuss single charged higgs boson production from @xmath0 annihilation at next generation linear colliders . they can be important to study the phenomenology of charged higgs bosons , especially when charged higgs bosons are too heavy to be produced by the pair production mechanism . cross sections for various single production processes are evaluated at the leading order . our analysis shows that in some parameter regions the phenomenology of charged higgs bosons can be explored even beyond the kinematic limit for pair production by using single production processes . cern - th/2001 - 025 + kek - th-743 + msuhep-01130
classical thermodynamics was originally designed to describe steam engines . it got its theoretical foundation in classical thermostatistics . now thermodynamics and the theory of phase transitions of homogeneous and large systems are some of the oldest and best established theories in physics . it may look surprising to add anything new to it . let me recapitulate what was told us since @xmath3 years : * thermodynamics addresses large homogeneous systems at equilibrium ( in the thermodynamic limit @xmath4 ) . * phase transitions are the positive zeros of the grand - canonical partition sum @xmath5 as function of @xmath6 ( yang - lee - singularities ) . of course these singularities indicate the break - down of the ( grand-)canonical ensemble . * micro and canonical ensembles are equivalent . * thermodynamics works with intensive variables @xmath7 . * unique legendre mapping @xmath8 . * heat only flows from hot to cold ( clausius ) . * second law only in infinite systems when the poincar recurrence time becomes infinite ( much larger than the age of the universe ( boltzmann ) ) . under these constraint only a tiny part of the real world of equilibrium systems can be treated . the ubiquitous non - homogeneous systems : nuclei , clusters , polymers , soft matter ( biological ) systems , but also the largest , astrophysical systems are not covered . even normal systems with short - range coupling at phase separations are inhomogeneous and can only be treated within conventional homogeneous thermodynamics ( e.g. in van - der - waals theory ) by bridging the unstable region of negative compressibility by a maxwell construction . thus even the original goal , for which thermodynamics was invented some @xmath9 years ago , the description of steam engines is only artificially solved . there is no ( grand-)canonical ensemble of phase separated and , consequently , inhomogeneous , configurations . this has a deep reason as i will discuss below : here the systems have a _ negative _ heat capacity @xmath10 ( resp . susceptibility ) . this , however , is impossible in the ( grand-)canonical theory where @xmath11 as was recently remarked by pitowsky@xcite : `` there is a schizophrenic attitude in the foundations of statistical mechanics . while boltzmann s view has been promoted as conceptional superior to that of gibbs , it is the canonical and not the microcanonical probability distribution that is extensively used in calculations . the claim being that the mathematical simplifications aside , all the ensembles are thermodynamically equivalent because as long we are dealing in systems containing a large number of molecules . '' . here , i will show that this is not only schizophrenic but is merely wrong in the most interesting situations . and that not only for small systems but also for the really large ones . the microcanonical ensemble is the ensemble ( manifold ) of all possible points in the @xmath12 dimensional phase space at the prescribed sharp energy @xmath13 : @xmath14 thermodynamics addresses the whole ensemble . it is ruled by the topology of the geometrical size @xmath15 , as expressed on boltzmann s epitaph : @xmath16 which is the most fundamental definition of the entropy @xmath17 . entropy and with it micro - canonical thermodynamics has a pure mechanical , geometrical foundation . no information theoretical formulation is needed . moreover , in contrast to the canonical entropy , @xmath18 is everywhere single valued and multiple differentiable . there is no need for extensivity , for concavity , for additivity , and no need for the thermodynamic limit . this is a great advantage of the geometric foundation of equilibrium statistics over the conventional definition of the boltzmann - gibbs ( bg ) canonical theory . however , addressing entropy to finite eventually small systems we will face a new problem with zermelo s objection against the monotonic rise of entropy , the second law . here the poincar recurrence time might be small and boltzmann s excuse does not work anymore . this is discussed elsewhere @xcite . the topology of the hessian of @xmath19 , the determinant of curvatures of @xmath19 , determines completely all kinds of phase transitions . this is evidently so , because @xmath20 is the weight of each energy in the canonical partition sum at given @xmath21 . consequently , at phase separation this has at least two maxima , the two phases . and in between two maxima there must be a minimum where the curvature of @xmath22 is positive . i.e. the positive curvature detects phase separation . this is of course true also in the case of several conserved control parameters e.g. @xmath23 , in the case of the @xmath24 potts - gas model on a two dimensional lattice of _ finite _ size of @xmath25 lattice points . here the hessian is [ @xmath26 are the eigenvalues = eigen - curvatures ] : @xmath27 the whole zoo of phase - transitions can be sharply seen and distinguished [ fig.([det ] ) ] . the curvature ( hessian ) of @xmath28 controls the phase transitions see ref.@xcite . what is the physics behind a positive curvature ? for a short - range force it is linked to the interphase surface tension c.f . in fig.[naprl0f ] and to a _ negative heat capacity_. this implys that _ heat can flow from cold to hot _ , c.f . fig[heat ] . in fig.[clusterfragmentation ] i show the simulation of a typical fragmentation transition of a system of @xmath29 sodium atoms interacting by realistic ( many - body ) forces . to compare with usual macroscopic conditions , the calculations were done at each energy using a volume @xmath30 such that the microcanonical pressure @xmath31atm . evidently , the convex region of @xmath32 is the most interesting region . self - gravitation leads to a non - extensive potential energy @xmath33 . no thermodynamic limit exists for @xmath34 and no canonical treatment makes sense . at negative total energies often these systems have a negative heat capacity . this was for a long time considered as an absurd situation within canonical statistical mechanics with its thermodynamic `` limit '' . however , within our geometric theory this is just a simple example of the pseudo - riemannian topology of the microcanonical entropy @xmath35 provided that high densities with their non - gravitational physics , like nuclear hydrogen burning , are excluded . we treated the various phases of a self - gravitating cloud of particles as function of the total energy and angular momentum , c.f . fig.[prl ] and the quoted prl - paper . clearly these are the most important constraint in astrophysics . in fig.[phasediagram ] the global phase diagram of a rotating , selfgravitating hydrogen cloud is given . entropy has a simple and elementary definition by the _ area _ @xmath36 of the microcanonical ensemble in the @xmath12 dim . phase space . canonical ensembles are not equivalent to the micro - ensemble in the most interesting situations : 1 . at phase - separation ( @xmath37heat engines ! ) , one gets inhomogeneities , and a negative heat capacity or some other negative susceptibility , consequently : 2 . heat can flow from cold to hot . 3 . phase transitions can be localized sharply and unambiguously in small classical or quantum systems , there is no need for finite size scaling to identify the transition . 4 . also really large self - gravitating systems can be addressed . entropy rises during the approach to equilibrium , @xmath38 , also for small mixing systems . i.e. the second law is valid even if the poincar recurrence time is not astronomically large @xcite . with this geometric foundation thermo - statistics applies not only to extensive systems but also to non - extensive ones which have no thermodynamic limit . more details are discussed in the references , see also my web - page http://www.hmi.de / people / gross/. the author is grateful to a.dellafiore , f.matera , m. pettini , and the members of the physics department of florence , but especially to s.ruffo , for many helpful discussions and their warm hospitality . he also acknowledges the financial support by the infn and the university of florence . d.h.e . ensemble probabilistic equilibrium and non - equilibrium thermodynamics without the thermodynamic limit . in andrei khrennikov , editor , _ foundations of probability and physics _ , number xiii in pq - qp : quantum probability , white noise analysis , pages 131146 , boston , october 2001 . acm , world scientific . d.h.e . second law in classical non - extensive systems . in d. sheehan , editor , _ proceedings of the first international conference on quantum limits to the second law _ , pages cond mat/0209467 , university of san diego , 2002 . d.h.e . thermo - statistics or topology of the microcanonical entropy surface . in t.dauxois , s.ruffo , e.arimondo , and m.wilkens , editors , _ dynamics and thermodynamics of systems with long range interactions _ , lecture notes in physics , 602 , pages 2145,cond mat/0206341 , heidelberg , 2002 . springer .
equilibrium statistics of hamiltonian systems is correctly described by the microcanonical ensemble . classically this is the manifold of all points in the @xmath0body phase space with the given total energy . due to boltzmann - planck s principle , @xmath1 , its geometrical size is related to the entropy @xmath2 . this definition does not invoke any information theory , no thermodynamic limit , no extensivity , and no homogeneity assumption . therefore , it describes the equilibrium statistics of extensive as well of non - extensive systems . due to this fact it is _ the fundamental _ definition of any classical equilibrium statistics . it addresses nuclei and astrophysical objects as well . @xmath2 is multiply differentiable everywhere , even at phase - transitions . all kind of phase transitions can be distinguished sharply and uniquely for even small systems . what is even more important , in contrast to the canonical theory , also the region of phase - space which corresponds to phase - separation is accessible , where the most interesting phenomena occur . no deformed q - entropy is needed for equilibrium . boltzmann - planck is the only appropriate statistics independent of whether the system is small or large , whether the system is ruled by short or long range forces . foundation of classical thermodynamics , non - extensive systems
the skyrme model is a classical non - linear field theory of interest both for itself and because it is a model of qcd in the limit of a large number of colors ( @xmath1 ) @xcite . it has been studied extensively in the baryon number one and number two sectors @xcite . recent results for baryon number two ( @xmath2 ) @xcite have shown that careful treatment of the non - linear nature of the equations and of quantum corrections ( finite @xmath1 effects ) , lead to a very good description of the nucleon - nucleon static potential . one finds one pion exchange , mid - range attraction , a tensor force and short range repulsion , all in substantial agreement with phenomenological potentials , and even with phase shifts @xcite . the next step is the consideration of non - static effects . here there has been less work @xcite and no systematic study . in this paper we present the analysis and results of a systematic study of the kinetic energy in the @xmath2 skyrmion system . we do so in the context of the product ansatz . this has the advantage of relative algebraic simplicity and of explicit identifiability of the collective coordinates . its main disadvantage is that it is a poor approximation to the @xmath2 exact solution at small skyrmion separation . for example , the product ansatz can not model the lowest energy @xmath2 static solution of the skyrme lagrangian in which the baryon density has toroidal symmetry and the identity of the individual skyrmions is completely lost . for the static nucleon - nucleon interaction , the product ansatz is particularly dangerous because that interaction is extracted by subtracting two large numbers to get a relatively small one , and even moderate errors in configurations get greatly magnified . that is why finding the mid - range attraction between nucleons waited for a full solution of the non - linear equations . for the non - static terms there is no corresponding subtraction and we might expect the product ansatz to be more reliable . however even here it is far from perfect . for example we find kinetic energy terms that are embarrassingly large , yet violate parity and particle interchange symmetry . these are artifacts of this ansatz , that has both these diseases . such terms would not be present in an exact calculation . nevertheless , as a first orientation , we calculate in the product ansatz . in the product ansatz , the @xmath2 skyrmion system is represented by two @xmath3 skyrmions separated by a distance @xmath4 . each of these two skyrmions can also be rotated in iso - spin space . for the static system , the energy of the @xmath2 state depends on @xmath4 and on the relative iso - spin orientation of the two skyrmions . to study non - static effects , we make each of these three collective vector coordinates , @xmath4 , and the iso - angle of each skyrmion , time dependent . now the kinetic energy can depend on any combination of the relative skyrmion velocity and the two angular velocities . clearly there are no terms in the kinetic energy linear in velocity . since the skyrme lagrangian is quadratic in time derivatives , the kinetic energy is only bilinear in the velocities . in principle all possible terms connecting velocities can occur in the kinetic energy , and since there are three vector velocities , the mass matrix will be nine by nine . the elements of this mass matrix depend on @xmath4 ( both its magnitude and direction ) , and the relative iso - spin orientation of the two skyrmions , and are calculated as integrals over the single skyrmion profile functions . there are many terms in the expansion of the kinetic energy in the relative iso - orientation . we find that terms have a steep hierarchy in size with relative iso - orientation , the terms that are independent of orientation being the largest . we also find , as mentioned above , that not all terms that violate parity and exchange symmetry are zero , due to the faults of the product ansatz . there are many terms in our expansion . in this paper we present the algebraic form for all of them but only discuss the largest ones . after the mass matrix has been constructed , it has to be inverted to pass from a velocity to a canonical momentum form so that a quantum mechanical hamiltonian can be constructed . we do not carry out that step here and only mention that attention to the requirements of matrix inversion are called for . we have recently demonstrated @xcite the role that the transformation from velocities to momenta plays in getting the correct sign for the nucleon - nucleon spin - orbit interaction . in sec . [ sec:2 ] we define notation , and substitute the @xmath2 product ansatz with time dependent collective coordinates in the skyrme lagrangian . we then develop the expansion of the terms in the kinetic energy . in sec . [ sec:3 ] we present the algebraic structure of the results of that expansion . section [ sec:4 ] discusses and presents graphically the behavior of some of the leading terms in our expansion . section [ sec:5 ] presents some conclusions . finally some technical details are given in the appendix . the main purpose of this paper is to present the full expansion of the kinetic energy both in systematic algebraic form , and as calculated results . we leave the use of these forms and results in particular physics problems to further publications . the skyrme model lagrange density can be written as @xcite @xmath5^{2 } \nonumber\\&&+\frac{f_{\pi}^2 m_{\pi}^2}{4 } [ \mbox{tr}(u+u^{\dagger})-4 ] , \label{eq : lag}\end{aligned}\ ] ] where @xmath6 is the left sugawara variable for the meson field @xmath7 which is a unitary @xmath8 matrix . there are only three parameters in the lagrangian , @xmath9 , the pion mass , @xmath10 , the pion decay constant , and @xmath11 , a dimensionless constant normally adjusted to give the correct nucleon - delta mass splitting . it is worth noting that the skyrme lagrangian is quadratic in time derivatives . the model has a topologically conserved quantum number that skyrme identified as the baryon number , @xmath12 . since the equations of motion derived from ( [ eq : lag ] ) are highly nonlinear , it is not possible , in general , to obtain exact analytical solutions . the standard time - independent solution to the classical field equations for @xmath3 is the defensive hedgehog , where the pion field points radially outward , @xmath13 the baryon number of this state is given by @xmath14 . this solution breaks translational invariance , as well as the @xmath15 spin - iso - spin symmetry . if we perform a global su(2 ) iso - rotation on the state , @xmath16 we obtain a state of the same energy . for the @xmath2 system , the product ansatz in which @xmath7 is written as a product of two single skyrmions is often used . the product ansatz is a poor approximation to the exact solution at short distances but it is expected to be valid when the two skyrmions are far apart . the advantage of using this ansatz is that , with it , the collective coordinates can be easily identified . we write @xmath7 for the skyrmion - skyrmion system as @xmath17 with @xmath18 and @xmath19 where @xmath20 and @xmath21 are both single skyrmion solutions , @xmath22 and @xmath23 are the positions of the two skyrmions with a relative distance r. finally @xmath24 and @xmath25 are the @xmath8 iso - rotation matrices for the two skyrmions . it is often convenient to work with the so - called sugawara variables of a @xmath8 matrix @xmath26 ( which can be @xmath7 , @xmath27,@xmath28 , @xmath20 , @xmath21 , @xmath29 or @xmath30 ) instead of the matrix itself . the left sugawara variable is defined as @xmath31 while the right one is @xmath32 the two sets of the sugawara variables are related , @xmath33 it can be shown that both @xmath34 and @xmath35 are traceless . we can , therefore , express these variables in terms of their three independent components @xmath36 and @xmath37 it is easy to show that for @xmath7 given in ( [ eq : prdtansatz ] ) : @xmath38 \tilde{u}_{1}^{\dagger}.\ ] ] using the trace formulas @xmath39 and @xmath40[{\vec}{\tau}\cdot{\vec}{c},{\vec}{\tau}\cdot{\vec}{d } ] \right ] = 8({\vec}{a}\cdot{\vec}{d})({\vec}{b}\cdot{\vec}{c})- 8({\vec}{a}\cdot{\vec}{c})({\vec}{b}\cdot{\vec}{d})$ ] , we find for the kinetic energy density @xmath41 where @xmath42 and @xmath43 if we use the hedgehog form for @xmath20 and @xmath21 , we find @xmath44 where @xmath45 and @xmath46 and @xmath47 . in the following we will use @xmath48 to represent the relative translational velocity . the quantities @xmath49 and @xmath50 are iso - rotational velocities expressed in terms of the right sugawara variables of the iso - rotation matrices . for isolated skyrmions , they are related to the spin operators of the individual skyrmions upon quantization by the relation @xmath51 where @xmath52 is the momenta of inertia for a single skyrmion . since the iso - rotation matrices @xmath29 and @xmath30 are global rotations , we have for the spatial components of the single skyrmion sugawara variables @xmath53 where @xmath54 and @xmath55 are the wigner @xmath56 functions for @xmath29 and @xmath30 and in terms of the components of the @xmath8 matrix @xmath57 , they are given by @xmath58 it is easy to verify that for single hedgehog skyrmions @xmath20 and @xmath21 , we have @xmath59 where @xmath60 s , @xmath61 s and @xmath62 s are scalar functions given by @xmath63 in terms of the hedgehog profile function @xmath64 as defined in ( [ eq : profile ] ) and its derivative @xmath65 . substituting eqs.([eq : hg1]-[eq : hg4 ] ) into ( [ eq : kt ] ) , we obtain for @xmath66 a sum of terms that are quadratic in the velocity vectors @xmath67 and @xmath50 with coefficients that depend on @xmath68 and the relative iso - rotation matrix @xmath69 . the kinetic energy is given by the integral over the kinetic energy density @xmath70 the three dimensional integral over all the terms in the expansion of @xmath66 can be reduced to a two dimensional integral by virtue of the axial symmetry around the axis connecting the centers of the two skyrmions . the integrands of these three dimensional integrals are all of the form @xmath71 where each @xmath72 is a component of the vector @xmath73 and @xmath74 is a non - negative integer . @xmath75 depends on @xmath76 and the angle @xmath77 between @xmath73 and @xmath4 . it is easy to see that @xmath78 where we use the notation @xmath79 it can also be shown that @xmath80 and similar relations for integrals over @xmath71 with @xmath81 . we use these relations to write the kinetic energy as a sum of scalar functions of the velocities , the unit vector @xmath82 of the relative separation of the two skyrmions and of the vector that specifies their relative iso - orientation . there have been a number of papers dealing with various selected terms in the kinetic energy . oka and odawara _ et al _ @xcite have studied the inertial parameters for translational motion only by neglecting the time - dependence of iso - rotations of the two skyrmions . riska and his co - workers @xcite have investigated terms in the kinetic energy that couple the translational velocity with the isorotational velocities and that contribute to the spin - orbit interaction when projected onto nucleon - nucleon states . a similar study was done by otofuji _ et al _ @xcite . however , our recent work @xcite has shown that a proper treatment of the spin - orbit interaction requires inversion of the mass matrix before projecting it onto nucleon states . this can not be done when terms that couple the translational velocity with the isorotational velocities are studied alone . in this work , we present a complete analysis of the kinetic energy within the product ansatz by expanding it in a set of rotational invariant quadratic functions of the translational and iso - rotational velocities and evaluating all coefficients in the expansion . the expansion can be conveniently divided into two groups . the first group , @xmath83 , contains terms involving a set of functions defined over the translational and iso - rotational velocity vectors . the second group , @xmath84 , is a linear combination of another set of functions defined over the three distinct doublets of the three velocity vectors . in both groups , we order terms by their irreducible representations under the relative skyrmion - skyrmion iso - orientation , @xmath85 . the terms quadratic in @xmath86 and @xmath87 we call @xmath88 ( for two ) , and those quartic we call @xmath64 ( for four ) . the @xmath88 s are the set of four dimensional harmonic polynomials of degree two and the @xmath64 s of degree four . they correspond to the symmetric representations of @xmath15 , @xmath89 and @xmath90 , respectively . since they are the symmetric representation , they have spin equal to iso - spin ( @xmath91 ) , with @xmath88 corresponding to @xmath92 and @xmath64 to @xmath93 . the symmetric representations of @xmath15 , @xmath94 , may be further distinguished by their grand - spin or @xmath95-spin , @xmath96 , where @xmath95 takes on the values @xmath97 . this then gives for @xmath88 a scalar term , @xmath98 , a vector term , @xmath99 , and a rank two term @xmath100 . for the forms quartic in @xmath85 , there are @xmath101 , @xmath102 , @xmath103 , @xmath104 , and @xmath105 . matrix elements of operators of the @xmath64-type vanish between two nucleon states . the detailed form for @xmath88 and @xmath64 in terms of @xmath85 are given in the appendix . using these expresssions we arrive at the following compact form for the expansion of the kinetic energy : @xmath106 where @xmath107 and @xmath108 \nonumber \\ & & + v^{(i)}_{1}({\vec}{t}_{1}\cdot{\vec}{t}_{2 } ) t^{(0 ) } + v^{(i)}_{2 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) t^{(0 ) } + v^{(i)}_{3 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2 } ) ] t^{(0 ) } \nonumber \\ & & + v^{(i)}_{4}({\vec}{t}_{1}\cdot{\vec}{t}_{2 } ) t^{(2)}(\hat{r},\hat{r } ) + v^{(i)}_{5 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) t^{(2)}(\hat{r},\hat{r } ) \nonumber \\ & & + v^{(i)}_{6 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2 } ) ] t^{(2)}(\hat{r},\hat{r } ) + v^{(i)}_{7 } ( { \vec}{t}_{1}\cdot\hat{r})t^{(2)}(\hat{r},\hat{r}\times{\vec}{t}_{2 } ) \nonumber \\ & & + v^{(i)}_{8 } ( { \vec}{t}_{2}\cdot\hat{r})t^{(2)}(\hat{r},\hat{r}\times{\vec}{t}_{1 } ) + v^{(i)}_{9 } ( { \vec}{t}_{1}\cdot\hat{r})t^{(2)}(\hat{r},{\vec}{t}_{2 } ) + v^{(i)}_{10 } ( { \vec}{t}_{2}\cdot\hat{r})t^{(2)}(\hat{r},{\vec}{t}_{1 } ) \nonumber \\ & & + v^{(i)}_{11 } t^{(2)}({\vec}{t}_{1},\hat{r}\times{\vec}{t}_{2 } ) + v^{(i)}_{12 } t^{(2)}({\vec}{t}_{2},\hat{r}\times{\vec}{t}_{1})+ v^{(i)}_{13 } t^{(2)}(\hat{r}\times{\vec}{t}_{1},\hat{r}\times{\vec}{t}_{2})+ \nonumber \\ & & + w^{(i)}_{1}({\vec}{t}_{1}\cdot{\vec}{t}_{2 } ) f^{(0 ) } + w^{(i)}_{2 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(0 ) } + w^{(i)}_{3 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2 } ) ] f^{(0 ) } \nonumber \\ & & + w^{(i)}_{4}({\vec}{t}_{1}\cdot{\vec}{t}_{2 } ) f^{(2)}(\hat{r},\hat{r } ) + w^{(i)}_{5 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(2)}(\hat{r},\hat{r } ) \nonumber \\ & & + w^{(i)}_{6 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2 } ) ] f^{(2)}(\hat{r},\hat{r } ) \nonumber \\ & & + w^{(i)}_{7 } ( { \vec}{t}_{1}\cdot\hat{r})f^{(2)}(\hat{r},\hat{r}\times{\vec}{t}_{2 } ) + w^{(i)}_{8 } ( { \vec}{t}_{2}\cdot\hat{r})f^{(2)}(\hat{r},\hat{r}\times{\vec}{t}_{1 } ) \nonumber \\ & & + w^{(i)}_{9 } ( { \vec}{t}_{1}\cdot\hat{r})f^{(2)}(\hat{r},{\vec}{t}_{2 } ) + w^{(i)}_{10 } ( { \vec}{t}_{2}\cdot\hat{r})f^{(2)}(\hat{r},{\vec}{t}_{1 } ) + w^{(i)}_{11 } f^{(2)}({\vec}{t}_{1},\hat{r}\times{\vec}{t}_{2})\nonumber \\ & & + w^{(i)}_{12 } f^{(2)}({\vec}{t}_{2},\hat{r}\times{\vec}{t}_{1})+ w^{(i)}_{13 } f^{(2)}(\hat{r}\times{\vec}{t}_{1},\hat{r}\times{\vec}{t}_{2})+ w^{(i)}_{14 } ( { \vec}{t}_{1}\cdot{\vec}{t}_{2 } ) f^{(4)}(\hat{r},\hat{r},\hat{r},\hat{r } ) \nonumber \\ & & + w^{(i)}_{15 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(4)}(\hat{r},\hat{r},\hat{r},\hat{r } ) + w^{(i)}_{16 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2 } ) ] f^{(4)}(\hat{r},\hat{r},\hat{r},\hat{r } ) \nonumber \\ & & + w^{(i)}_{17 } ( { \vec}{t}_{1}\cdot\hat{r } ) f^{(4)}(\hat{r},\hat{r},\hat{r } , { \vec}{t}_{2 } ) + w^{(i)}_{18 } ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(4)}(\hat{r},\hat{r},\hat{r } , { \vec}{t}_{1})\nonumber \\ & & + w^{(i)}_{19 } ( { \vec}{t}_{1}\cdot\hat{r } ) f^{(4)}(\hat{r},\hat{r},\hat{r},\hat{r}\times { \vec}{t}_{2 } ) + w^{(i)}_{20}({\vec}{t}_{2}\cdot\hat{r } ) f^{(4)}(\hat{r},\hat{r},\hat{r},\hat{r}\times{\vec}{t}_{1 } ) \nonumber \\ & & + w^{(i)}_{21}f^{(4 ) } ( \hat{r},\hat{r},{\vec}{t}_{1},\hat{r}\times { \vec}{t}_{2 } ) + w^{(i)}_{22}f^{(4 ) } ( \hat{r},\hat{r},{\vec}{t}_{2},\hat{r}\times { \vec}{t}_{1 } ) \nonumber \\ & & + w^{(i)}_{23}f^{(4)}(\hat{r},\hat{r},\hat{r}\times{\vec}{t}_{1 } , \hat{r}\times{\vec}{t}_{2 } ) \nonumber \\ & & + \phi^{(i)}_{1 } ( { \vec}{t}_{1}\cdot{\vec}{t}_{2 } ) t^{(1)}(\hat{r } ) + \phi^{(i)}_{2 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) t^{(1)}(\hat{r } ) + \phi^{(i)}_{3 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2})]t^{(1)}(\hat{r } ) \nonumber \\ & & + \phi^{(i)}_{4 } ( { \vec}{t}_{1}\cdot\hat{r } ) t^{(1)}({\vec}{t}_{2 } ) + \phi^{(i)}_{5 } ( { \vec}{t}_{2}\cdot\hat{r } ) t^{(1)}({\vec}{t}_{1 } ) \nonumber \\ & & + \phi^{(i)}_{6 } ( { \vec}{t}_{1}\cdot\hat{r } ) t^{(1)}(\hat{r}\times{\vec}{t}_{2 } ) + \phi^{(i)}_{7 } ( { \vec}{t}_{2}\cdot\hat{r } ) t^{(1)}(\hat{r}\times{\vec}{t}_{1 } ) \nonumber \\ & & + \eta^{(i)}_{1 } ( { \vec}{t}_{1}\cdot{\vec}{t}_{2 } ) f^{(1)}(\hat{r } ) + \eta^{(i)}_{2 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(1)}(\hat{r } ) + \eta^{(i)}_{3 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2})]f^{(1)}(\hat{r } ) \nonumber \\ & & + \eta^{(i)}_{4 } ( { \vec}{t}_{1}\cdot\hat{r } ) f^{(1)}({\vec}{t}_{2 } ) + \eta^{(i)}_{5 } ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(1)}({\vec}{t}_{1 } ) + \eta^{(i)}_{6 } ( { \vec}{t}_{1}\cdot\hat{r } ) f^{(1)}(\hat{r}\times{\vec}{t}_{2 } ) \nonumber \\ & & + \eta^{(i)}_{7 } ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(1)}(\hat{r}\times{\vec}{t}_{1 } ) + \eta^{(i)}_{8 } ( { \vec}{t}_{1}\cdot{\vec}{t}_{2 } ) f^{(3)}(\hat{r},\hat{r},\hat{r } ) \nonumber \\ & & + \eta^{(i)}_{9 } ( { \vec}{t}_{1}\cdot\hat{r } ) ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(3)}(\hat{r},\hat{r},\hat{r } ) + \eta^{(i)}_{10 } [ \hat{r}\cdot({\vec}{t}_{1}\times{\vec}{t}_{2 } ) ] f^{(3)}(\hat{r},\hat{r},\hat{r } ) \nonumber \\ & & + \eta^{(i)}_{11 } ( { \vec}{t}_{1}\cdot\hat{r } ) f^{(3)}(\hat{r},\hat{r},{\vec}{t}_{2 } ) + \eta^{(i)}_{12 } ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(3)}(\hat{r},\hat{r},{\vec}{t}_{1 } ) \nonumber \\ & & + \eta^{(i)}_{13 } ( { \vec}{t}_{1}\cdot\hat{r } ) f^{(3)}(\hat{r},\hat{r},\hat{r}\times{\vec}{t}_{2 } ) + \eta^{(i)}_{14 } ( { \vec}{t}_{2}\cdot\hat{r } ) f^{(3)}(\hat{r},\hat{r},\hat{r}\times{\vec}{t}_{1 } ) \nonumber \\ & & + \eta^{(i)}_{15 } f^{(3)}(\hat{r},{\vec}{t}_{1},{\vec}{t}_{2 } ) + \eta^{(i)}_{16 } f^{(3)}(\hat{r},{\vec}{t}_{1},\hat{r}\times{\vec}{t}_{2 } ) \nonumber \\ & & + \eta^{(i)}_{17 } f^{(3)}(\hat{r},{\vec}{t}_{2},\hat{r}\times{\vec}{t}_{1 } ) + \eta^{(i)}_{18 } f^{(3)}(\hat{r},\hat{r}\times{\vec}{t}_{1},\hat{r}\times{\vec}{t}_{2}).\end{aligned}\ ] ] for @xmath83 , we take @xmath109 and for @xmath84 , we take @xmath110 careful readers may wonder why a term such as @xmath111 is not present in the expansion above . it is because this term can be written as a linear combination of other terms in the expansion . we eliminate this and other terms that are linearly dependent on the invariant functions used in the expansion . we have used roman characters for coefficients of those terms that are independent of @xmath112 and greek letters for those terms that involve @xmath112 . we find that for the pair @xmath113 and the pair @xmath114 all roman letter coefficients have the same sign for the two terms in each pair while all the greek letter coefficients have the opposite sign . this implies that terms that violate parity are also anti - symmetric under the exchange of the two skyrmions . in the expansion above , we have included parity - violating terms that are artifacts of the product ansatz . of course in an exact calculation , these terms will not appear . in this section we present the @xmath115 dependence of the leading terms in the expansion of the kinetic energy , @xmath116 and @xmath117 , of the last section . the kinetic energy is a general bilinear form in the three velocity vectors , the relative translational velocity , and the angular velocity of each of the individual skyrmions . the first of these we call @xmath118 , the angular velocities we call @xmath119 . the parameters of the skyrme model we use are taken from @xcite and are listed in table 1 . .skyrme model parameters [ cols="^,^,^",options="header " , ] the coefficients for terms that are quadratic in @xmath118 are given in units of mev since @xmath118 is given in units of c , the speed of light . the coefficients for terms that are linear in @xmath118 and @xmath120 are given in units of @xmath121 since @xmath122 has the dimension of 1/fm . it follows that the coefficients for terms that are quadratic in @xmath120 have units of @xmath123 . we first consider the terms in @xmath124 in which @xmath125 , the relative velocity between the skyrmions . we consider the dependence on @xmath115 of each of the coefficients of the invariants in ( [ eq : kexp ] ) . the first terms , @xmath126 and @xmath127 , do not depend on the relative iso - orientation , @xmath85 . we find that @xmath126 is quite large while @xmath127 is small . the @xmath126 term goes over to a constant at large @xmath115 , since it becomes the ordinary translational kinetic energy of the single separated skyrmions . in that case , and in our units and conventions , @xmath126 is just @xmath128 the mass of a single skyrmion . we show @xmath126 plotted as a function of @xmath115 in fig . 1 . the solid line is the calculation with the physical pion mass and the dotted line with zero pion mass . the effect of the pion mass is relatively small , and only significant inside 1 fm , where we should distrust the product ansatz . we find that @xmath129 and @xmath130 are quite small , @xmath131 and @xmath132 of moderate size and @xmath133 and @xmath134 large at small to medium values of @xmath115 ( see figs . @xmath135 is odd under parity and interchange of the two skyrmions . no such terms should appear in a correct calculation , and its large size out to medium @xmath115 is a measure of the limitations of the product ansatz . all further terms in the expansion of @xmath124 with @xmath125 are quite small outside @xmath136 fm , with the @xmath137 terms particularly small . these are @xmath64 terms ( quartic in @xmath85 ) and their small size is a manifestation of the decreasing size of terms with increasing iso - spin complexity . the @xmath138 and @xmath139 terms are all identically zero because @xmath140 is in fact invariant under the transformation @xmath141 . next we consider @xmath142 in @xmath143 . the results for @xmath144 are the same up to terms that are not symmetric under interchange of the two skyrmions . these terms violate parity . the term in @xmath145 is a pure iso - kinetic energy . since for an individual skyrmion an iso - rotation is equivalent to a spin rotation , this term goes over to the kinetic energy of a freely rotating skyrmion for large @xmath115 . there @xmath145 is equal to twice the moment of inertia @xmath146 . all this is seen in fig . 5 where we also see that there is more sensitivity to the pion mass than in the @xmath125 terms , but mostly at small distances where the product ansatz is not to be trusted . we find that @xmath147 , @xmath148 , and @xmath149 are all quite small except at very small @xmath115 . @xmath150 , @xmath151 , @xmath152 , and @xmath153 are all of moderate size beyond 1 fm , and show interesting @xmath115 dependence and some sensitivity to the pion mass ( see figs . @xmath154 ( see fig . 8) is a wrong parity term that is comparable with @xmath150 , @xmath151 , @xmath152 , and @xmath153 . the next term that is significant beyond 1 fm is @xmath155 . it is one of the few quartic terms in @xmath85 that is large ( see fig . it is also remarkably sensitive to the inclusion of a pion mass term . the remaining quartic terms are small as are the greek letter terms . the largest of these are @xmath156 and @xmath157 , which we show in fig . we now turn to the terms in @xmath117 . in these terms two kinds of velocities are coupled . we consider first @xmath158 and @xmath159 . again the contribution with @xmath160 can be obtained from the first by exchanging one and two with the understanding that terms that violate parity are also odd under the interchange . we find that @xmath161 and @xmath162 are zero . these terms would violate parity since @xmath118 and @xmath119 behave oppositely under parity . we find that @xmath163 is relatively large ( see fig . it is the term principally responsible for iso - spin independent spin - orbit coupling . it has been calculated before @xcite but our result is about as twice as large as reported there . more importantly even though we obtain the same sign for @xmath163 as reported in @xcite , a careful conversion from a velocity to a canonical momentum form leads to the correct sign for the spin - orbit interaction , as we have shown in our recent work @xcite . @xmath164 , @xmath165 and @xmath166 are very small , but @xmath167 and @xmath168 are large at small @xmath115 , and parity violating ( see fig . @xmath169 , @xmath170 , @xmath171 , @xmath172 and @xmath173 are moderately large , but not as large as @xmath163 , and contribute to the iso - spin dependent spin - orbit interaction ( see figs . 1315 ) . @xmath174 , @xmath175 and @xmath176 are very large parity violating terms that decay slowly with @xmath115 ( see figs . 1617 ) . they are troublesome . all of the @xmath177 terms ( quartic in @xmath85 ) are essentially zero outside 1 fm . of the greek letter terms , @xmath178 , @xmath179 and @xmath180 are parity conserving and large , while @xmath181 and @xmath182 are large , slowly decreasing with @xmath115 and parity violating ( see figs . 1820 ) . all of the @xmath183 terms ( quartic in @xmath85 ) are very small outside of 1 fm . finally we consider @xmath117 with @xmath184 and @xmath185 . we find @xmath186 , @xmath187 , @xmath188 , @xmath189 , @xmath190 , @xmath191 , @xmath192 , @xmath193 , @xmath194 , and @xmath195 all to be very small . we find @xmath196 , @xmath197 , @xmath198 and @xmath199 to be large ( see figs . 2122 , @xmath198 is not plotted since it is equal to @xmath200 as expected from permutation symmetry ) . they also decay very slowly for large @xmath115 . analysis shows that they go like @xmath0 for large @xmath115 in the case of zero pion mass . the remaining terms in @xmath201 and all the terms in @xmath202 ( quartic in @xmath85 ) are small outside 1 fm . @xmath203 , @xmath204 and @xmath205 are very large and exhibit considerable sensitivity to pion mass ( see fig . 23 , @xmath205 is not plotted since it is equal to @xmath206 as expected from permutation symmetry ) . and @xmath207 are large and parity violating ( see fig . 24 , only @xmath208 is plotted since @xmath209 , thus violating permutation symmetry as well ) . again all the quartic terms , @xmath210 , are small . in this paper we present a systematic development of the kinetic energy of the @xmath2 skyrmion system starting from the product ansatz . although this ansatz has serious shortcomings , it provides a useful starting point for a first orientation to the velocity dependent terms in the skyrme energy . it is particularly easy to identify the collective degrees of freedom in the product ansatz and therefore to make them time dependent . furthermore the problems of the product ansatz in calculating the static energy of the @xmath2 system , namely the subtraction of large numbers to get a small one , are not present in the kinetic energy . the full kinetic energy of the @xmath2 skyrmion system is a rich and complex expression . there are three dynamical velocities that enter , the relative velocity of the two skyrmions and the rotational velocity of each . we have not considered the trivial overall translational velocity of the entire system . the kinetic energy is bilinear in the velocities and all possible couplings among the three velocity vectors enter , so that the mass matrix is nine by nine . the kinetic energy is a scalar function of the velocities , the relative separation vector of the two skyrmions and of the vector that specifies their relative iso - orientation . it is convenient to classify the dependence on iso - orientation in terms of irreducible tensors of definite degree in the relative orientation , @xmath85 . call that degree @xmath211 . ( it is in fact the @xmath15 quantum number , see appendix ) . we find terms with @xmath212 . the terms with @xmath213 correspond to no iso - orientational dependence and lead to iso - spin independent contributions to the kinetic energy of two nucleons . the terms with @xmath214 lead to kinetic energy terms for two nucleons that have an iso - spin dependence of the form @xmath215 we find that , in general , the @xmath214 terms are smaller for the two skyrmion system than the @xmath213 terms . the @xmath216 terms are smaller still and have vanishing matrix element in the two nucleon system , but will couple with deltas . we find that there are some non - zero terms in the expansion of the kinetic energy that are not even under parity and are not symmetric under interchange of the two skyrmions . these terms are pathological artifacts of the product ansatz that would not be present in an exact calculation . some of these terms are quite large and fall off very slowly with the separation of the skyrmions , particularly in the case of zero pion mass . these large terms are particularly worrisome and give an ominous sense of the limits of the product ansatz calculation . our major purpose is to give a full and systematic discussion of the two skyrmion kinetic energy to serve as a foundation for any discussion of non - static terms in the two skyrmion and two nucleon interaction . here we only present the full algebraic development of the kinetic energy and some result for the largest terms with their dependence on the relative separation of the skyrmions . there are many applications of these results that come to mind . for example the leading non - static contribution to the energy of two nucleons is the spin - orbit interaction . it arises from the terms in the kinetic energy that couple the relative velocity and the total rotational velocity . to go from these terms to a nucleon - nucleon spin - orbit interaction in a hamiltonian require that the mass matrix be inverted before projecting on nucleon states . we have recently shown that doing this yields an iso - spin independent spin - orbit interaction of the correct magnitude and sign as well as explaining the relatively small size of the iso - spin dependent spin - orbit interaction @xcite . there are many more non - static terms in the nuclear force that invite study . there is also the opportunity to go beyond the product ansatz . here the problem of defining collective variables comes up . we are studying this difficult issue . a starting point , pursued with considerable success in the static energy studies , is to start with the kinetic energy in the two skyrmion channels with reflection symmetry . this is easily done in the product ansatz and we are studying the advantages of this approach for an exact treatment . in summary , we have studied the kinetic energy of the @xmath2 skyrmion system in the product ansatz in a systematic expansion in the collective velocitites . this is a first step in adding non - static corrections to the recent success for the static nucleon - nucleon interaction @xcite calculated in an approach based in non - perturbative qcd . this research was supported in part by a grant from the u. s. national science foundation . the expansion of the kinetic energy is a sum of terms quadratic in the translational and isorotational velocities . the coefficients depend only on the relative isorotation @xmath69 . since @xmath217 is unitary , the @xmath218 s satisfy @xmath219 the operators @xmath220 form an @xmath15 algebra . the @xmath221 s generate an @xmath222 subalgebra . the casimir operator of @xmath15 is @xmath223 and its eigenvalue is @xmath224 . the casimir operator of the @xmath222 subalgebra is @xmath225 with eigenvalue @xmath226 . it is easy to show that the following polynomials in the four components of @xmath85 , spherical harmonics of @xmath15 , carry irreducible representations of the group chain @xmath227 : the @xmath236 polynomials are quadratic in @xmath85 and carry the irreducible representations @xmath237 of @xmath227 . they are linear combinations of the matrix elements of the wigner function @xmath238 . the @xmath239 polynomials are quartic in @xmath85 and carry the irreducible representations @xmath240 of @xmath227 . are linear combinations of the matrix elements of the wigner function @xmath241 and their matrix elements between nucleon - nucleon states are therefore zero . 99 t. h. r. skyrme , proc . london * 260 * , 127 ( 1961 ) ; * 262 * , 237 ( 1961 ) ; nucl . phys . * 31 * , 556 ( 1962 ) . i. zahed and g. e. brown , phys . rep . * 142 * , 1 ( 1986 ) ; _ chiral solitons _ , edited by k. f. liu ( world scientific , singapore , 1987 ) and references therein . a. jackson , a. d. jackson and v. pasquier , nucl . phys . * a432 * , 567 ( 1985 ) ; r. vinh mau , m. lacombe , b. loiseau , w. n. nottingham and p. lisboa , phys . lett . * b150 * , 259 ( 1985 ) ; t. s. walhout and j. wambach , phys . * 67 * , 314 ( 1991 ) ; t. s. walhout and j. wambach , int . j. mod e1 * , 665 ( 1992 ) ; n. j. snyderman , in : _ baryons as skyrme solitons 1992 _ , edited by g. holtzwarth ( world scientific , singapore , 1993 ) r. d. amado , r. bijker and m. oka , phys . lett . * 58 * , 654 ( 1987 ) ; m. oka , r. bijker , a. bulgac and r. d. amado , phys . rev . * c36 * , 1727 ( 1987 ) . n. r. walet , r. a. amado and a. hosaka , phys . . lett . * 68 * , 3849 ( 1992 ) ; n. r. walet and r. d. amado , phys * c47 * , 498 ( 1993 ) . n. r. walet , phys . rev . * c * , ( submitted , 1993 ) . m. oka , phys . b175 * , 15 ( 1986 ) . h. odawara , o. morimatsu and k. yazaki , phys . lett . * b175 * , 115 ( 1986 ) . d. o. riska and k. dannbom , physica scripta * 37 * , 7 ( 1988 ) . t. otofuji , s. saito , m. yasuno , h. kanada , and r. seki , phys . * b205 * , 145 ( 1988 ) . r. d. amado , b. shao and n. r. walet , preprint upr-0127mt ( 1993 ) . g. s. adkins , c. r. nappi and e. witten , nucl . phys . * b228 * , 552 ( 1983 ) .
the kinetic energy of the skyrmion - skyrmion system is calculated in the product ansatz . the energy is expanded in a set of rotationally invariant quadratic functions of translational and iso - rotational velocities in the product ansatz and all inertial parameters are evaluated . most of the iso - spin dependent parameters are small at separations larger than 1 fm . however , some parameters of the coupling between the iso - rotational velocities of the skyrmions decay slowly , as @xmath0 , when zero pion mass is used . otherwise we find that the results are not sensitive to the addition of the pion mass term . + pacs : 21.30.+y , 13.75.cs , 12.40.-y , 11.10.lm + * preprint upr0128mt * + july 1993 * inertial parameters of the skyrmion - skyrmion system with the product ansatz * + * bin shao , niels r. walet , and r. d. amado *
the efimov effect has been predicted by v. efimov in a series of papers @xcite and experimentally confirmed more than 35 years after its prediction @xcite . at present there is an intense experimental activity @xcite aiming at understanding the behavior of few - body systems close to the unitary limit in which the two - body scattering length @xmath0 diverges . around this limit the systems show universal behavior ; very different systems such as atomic or nuclear systems present similar features . the study of these characteristics is a very active field of research nowadays . from a theoretical point of view the behavior of a few - body system in the limit of large scattering length can be formulated in the framework of the renormalization group using an effective field theory ( eft ) @xcite ( for recent reviews see refs . @xcite ) . using this language , one of the main striking properties of three identical bosons in the unitary limit is the discrete scaling invariance ( dsi ) shown by the spectrum : an infinite series of bound states ( trimers ) appears distributed geometrically with accumulation point at zero energy . the ratio of binding energies for two consecutive states is @xmath2 , with the universal number @xmath3 . explicitly , the total angular momentum @xmath4 spectrum of three identical bosons in the zero - range limit can be described by the efimov radial law @xmath5 [ eq : energyzr ] the main ingredients in these equations are the universal function @xmath6 and the binding momentum @xmath7 , called the three - body parameter , defining the energy @xmath8 of level @xmath9 at the unitary limit , @xmath10 being the boson mass . dsi manifests from the fact that the function @xmath6 is the same for all @xmath11 levels . furthermore , the spectrum described by these equations is not bounded from below , this characteristic is known as the thomas collapse @xcite . the determination of the universal function @xmath6 in the interval @xmath12 limited by the two- and three - cluster continuum can be obtained by solving the skorniakov - ter - martirosian ( stm ) equation or equivalently by using effective field theory ( eft ) as reported in ref . this equation describes the @xmath4 state of a three - boson system in the zero - range limit . to avoid the thomas collapse it is a common practice to introduce a cutoff in the solution of the stm equation and the universal function @xmath6 is computed looking at the second or even third excited state @xcite where the cutoff effects are negligible . a parametrization of it can be found in ref . in the same way potential models can be used to solve the schrdinger equation looking at the high part of the spectrum , where finite - range effects are negligible . the extension of the zero - range theory to four bosons has been discussed in refs . the main conclusion of these works is that the four - boson spectrum presents a two - level tree structure . for each three - body level @xmath13 there are two four - body states ( tetramers ) , one deep ( @xmath14 ) and one shallow ( @xmath15 ) , with binding energies @xmath16 . the universal ratios of these binding energies in the unitary limit have been calculated in ref . @xcite and they are @xmath17 and @xmath18 . in the present work we would like to analyze the complete interval between the four - body continuum and the dimer - dimer thresholds in order to extend eq.([eq : energyzr ] ) to the four - boson system and , if possible , to determine the corresponding universal function . the present study is done using potential models with variable strength . in this way a path to reach the unitary limit is defined . it has been recently shown that a two - parameter potential captures the essential ingredients of the few - boson dynamics close to the unitary limit @xcite . accordingly , here we define two different potentials , a local gaussian and a nonlocal gaussian used to solve the schrdinger equation along the path . for the lowest states finite - range effects are appreciable . for this case it is possible to define level functions that absorb those effects and can be used to estimate the spectrum of a real system close to the unitary limit . for higher states finite - range effects can be neglected and the solution tends to the zero - range limit therefore the level functions tend to the universal zero - range function . in order to illustrate the procedure we first analyse the two - body system and then the analysis is extended to the three- and four - boson systems . the paper is organized as follows . in sec.ii the particular path used to reach the unitary limit is studied in the two - body system . in sec . iii and iv the spectra of the three- and four - boson systems are computed using the two potential models , local and nonlocal . from the results a zero - range equation for the four - boson system is proposed . the conclusion and perspectives are given in the last section . in the two - body system the @xmath4 spectrum within the zero - range theory can be defined through the relations @xmath19 [ eq : energy2bzr ] the first relation establishes that there is a bound state ( @xmath20 ) or a virtual state ( @xmath21 ) with binding energy @xmath1 fixed by the scattering length @xmath0 . at positive energies , @xmath22 , the @xmath23-wave phase - shift @xmath24 is determined by the scattering length as well . accordingly , the scattering length @xmath0 emerges as a control parameter in terms of which the observables as the cross section or mean square radius can be computed . the zero - range theory describes the extreme situation in which the two particles are always outside the interaction range . if the two - body quantum system , interacting through a short - range potential , has a shallow state , there is a big probability of finding the particles outside the interaction range . in fact , when @xmath1 is very small or , equivalently , @xmath25 , with @xmath26 the interaction range , the two - body wave function has a very long tail and the two particles have a large probability of being at relative distances greater than @xmath26 . when a shallow state is present , the scattering length verifies @xmath27 too , and @xmath28 , with eq.([eq : energy2bzra ] ) approximately fulfilled . the extension of eq.([eq : energy2bzr ] ) in the case of finite - range interactions and at low energies is @xmath29 [ eq : energy2bfr ] with @xmath30 being the effective range . in the case of shallow states the second equation can be used to relate the effective range to @xmath31 , @xmath32 from which we obtain the relation @xmath33 with @xmath34 . the above discussion stressed the fact that in the low energy limit or large scattering length limit the dynamics of the two - body system depends on two parameters : the scattering length and the effective range @xmath30 ( or the length @xmath35 ) . the scaling limit is defined by @xmath36 whereas in the unitary limit @xmath37 and @xmath38 . in the first case , for each value of @xmath0 the two - body energy is determined by the zero - range condition @xmath39 . when @xmath40 , the unitary limit can be reached by different paths determined by the functional relation @xmath41 . having in mind that we intend to study the structure of few - boson systems , we construct a two - parameter potential able to reproduce the minimal information given by one specific set of @xmath0 and @xmath31 values . then the potential strength can be varied in order to reach the unitary limit . we define local ( given in coordinate space ) and nonlocal ( given in momentum space ) gaussian potentials @xmath42 with the strengths @xmath43 , @xmath44 and the ranges @xmath26 , @xmath45 determined to describe the particular value of @xmath0 and @xmath31 of a two - boson system . the parameter @xmath46 can be varied in order to reach the unitary limit . with this procedure the functional relation @xmath41 is determined . it should be noted that with the potentials defined above the lengths , momenta and energy scale with @xmath26 , @xmath47 and @xmath48 ( or @xmath49 ) , respectively . accordingly , the local gaussian defines a particular path to the unitary limit that encompasses all local gaussians and the same for the nonlocal one . in particular , the values of the effective range and strength at unitary are given in table [ tab : tab1 ] for the cases in which there are @xmath11 bound states in the two - body system . note that the nonlocal potential , being rank - one separable potential , supports only one bound state at most . .universal values of the effective range ( in units of @xmath26 or @xmath45 ) and potential strength ( in units of @xmath48 or @xmath50 ) for local and nonlocal gaussians at the unitary limit for different bound state numbers @xmath11 . [ cols="^ , < , < , < , < " , ] [ tab : tab1 ] in the following we consider the @xmath51 case , however the other cases can be analyzed in a similar way . we are interested in the functional relation @xmath52 as @xmath53 . defining @xmath54 the value of the effective range at the unitary limit , from eq.([eq : reff ] ) we can define @xmath55 with @xmath56 . moreover , the quantity @xmath57 can be expanded around the unitary limit as @xmath58 inserting the above expansion in eq.([eq : reffru ] ) the effective range expansion in terms of the inverse of the scattering length becomes @xmath59 if the length @xmath35 were constant along the path , the coefficient @xmath60 would be @xmath61 and @xmath62 as well as higher terms . differences from these values indicate a non constant behavior of @xmath35 . for example the study of the van der waals ( vdw ) potential shows a quadratic relation between @xmath30 and the inverse of the scattering length . explicitly it results ( see ref . @xcite and references therein ) @xmath63 with @xmath64 . therefore the coefficients in eq.([eq : reffx ] ) are @xmath65 and @xmath66 . in the case of the lg and nlg potentials the behavior is almost linear ( @xmath67 ) , with the coefficient @xmath68 and @xmath69 , respectively . the relation @xmath41 can be analyzed starting from the following definition @xmath70 defining @xmath56 and @xmath71 , and considering the expansion of @xmath57 given above , this relation , at first order in @xmath37 , results @xmath72 in fig . [ fig : fig2 ] the quantity @xmath73 is shown for different potential models around the unitary limit . the results of the different potential models collapse in the curve given by eq.([eq : unixy ] ) showing that up to first order the relation @xmath74 is well verified . in the study we have included the lm2m2 helium - helium interaction from aziz @xcite . we can conclude that eq.([eq : unixy ] ) can be seen as a universal relation describing the path to the unitary limit fixed by the relation @xmath75 . potentials with variable strength follow this path with reasonable accuracy , however first order corrections could be of the order of a few percent for the lg potential and up to @xmath76 in the case of the nlg potential . and @xmath0 ( in units of @xmath77 ) for the different potential models . all the computed values collapse on the curve @xmath78 . ] the analysis of the two - boson system indicates that a two - parameter potential as local or non - local gaussian can be used to study the low - energy dynamics around the unitary limit . we will extend the analysis to the three- and four - boson systems . the numerical results for the local gaussian are obtained solving the schrdinger equation in the coordinate space framework using the hyperspherical harmonic expansion @xcite while the predictions for the non - local gaussian are obtained solving faddeev - yakubovsky ( fy ) @xcite or alt - grassberger - sandhas ( ags ) equations @xcite using momentum - space methods from refs . @xcite . the @xmath4 spectrum of three identical bosons in the zero - range limit can be described by the efimov radial law given in eq.([eq : energyzr ] ) . in the case of a system with a finite range interaction , this equation can be seen as describing the asymptotic spectrum of the three - boson system close to the unitary limit . in fact , the solution of the schrdinger equation can be used to determine the universal function @xmath6 looking at the excited states of the spectrum as , for these states , finite - range effects are negligible . in this context the description of few - body systems with potential models close to the unitary limit can be seen as a particular regularization scheme . accordingly it is possible to modify the efimov radial law as ( see refs . @xcite ) @xmath79 [ eq : energyfr3 ] where @xmath80 is the dimer binding energy for positive values of @xmath0 whereas for negative values it is the two - body virtual state energy . modifications at the three - body level are introduced by the parameters @xmath81 who absorb the scaling factor @xmath82 defining the energy of level @xmath11 at the unitary limit , @xmath83 . furthermore the finite - range character of the interaction slightly modifies the ratio @xmath84 from its universal value of @xmath85 . the main modification in the above equations is the introduction of the level function @xmath86 . for the ground state ( @xmath51 ) it could be very different from the zero - range function @xmath6 . as we will see below the differences are much reduced considering the first excited state ( @xmath87 ) and , starting from @xmath88 , both functions almost coincide . the level function can be calculated using the corresponding solutions of the schrdinger equation as @xmath89 it should be noticed that for @xmath90 this function depends on the particular potential used to calculate the spectrum or , in the case of the stm equation , the cutoff . it depends also on the particular path selected to reach the unitary limit , for example , potentials with variable strength as discussed in the previous section . however , as shown in ref . @xcite , following this particular path different potentials do not produce too much spread in @xmath91 and @xmath92 . furthermore the lg potential defines a unique gaussian function @xmath86 for each level @xmath11 independent of the range , @xmath26 , of the potential . in particular , for the first two levels @xmath90 , the binding momenta at the unitary limit are @xmath93 and @xmath94 and the ratio @xmath95 . these values have been obtained with the potential acting only in @xmath23-waves , they are slightly different when the lg potential is taken to act in all waves ( see for example refs . @xcite ) . also the nlg potential acting in @xmath23-wave defines a unique nonlocal gaussian function with the following universal ratios @xmath96 , @xmath97 and @xmath98 . from the above discussion the following picture emerges : a two - parameter potential as the local or nonlocal gaussians can be used to construct level functions for each level @xmath11 of the three - boson system . for @xmath90 these functions are different from the zero - range universal function and are also different among themselves . for @xmath88 they converge to the zero - range universal function showing a universal behavior . in order to analyze this fact quantitatively , in fig . [ fig : fig3 ] the level functions @xmath91 and @xmath92 are shown for the case of the lg , nlg and lm2m2 potentials . the trend discussed above is visible in the figure , for the ground state the level functions @xmath91 spread in a narrow band and are rotated with respect to the universal function @xmath6 , given by the red solid curve @xcite . in the case of the first excited state the level functions @xmath92 spread also in a very narrow band very close to @xmath6 . the level function @xmath99 calculated using the nlg third excited state ( red diamonds ) completely overlaps with the zero - range universal function . for different levels and potentials . the zero - range universal function @xmath6 is shown as the red solid curve . ] the efimov radial law is a one - parameter equation . the knowledge of the universal function @xmath6 allows for a complete determination of the spectrum after assigning a value to @xmath7 ( or to one of the energies @xmath13 ) . eq.([eq : energyfr3 ] ) applies mostly to @xmath90 and works slightly differently . in first place it is necessary to calculate the level functions for the ground and first excited state . this can be done using for example the lg or nlg potentials . at this point the equation is a one - parameter equation and the @xmath51 and @xmath87 spectrum can be completely determined from the knowledge of one energy . one can argue that if we use eq.([eq : energyfr3 ] ) to describe a particular system , the respective potential could be used to compute the level function @xmath91 or @xmath92 by varying the strength in order to reach the unitary limit . however , as it is shown in fig . [ fig : fig3 ] , close to the unitary limit the three - boson system has universal behavior and therefore a two - parameter potential captures the essential ingredients of the dynamics absorbing finite range effects . as an example we can use @xmath100 computed using the lg or nlg potentials to estimate the three - body parameter @xmath101 of a system composed by three @xmath102he atoms . as given in ref . @xcite the result using the lg potential is @xmath103a@xmath104 ( a@xmath105 is the bohr radius ) . using the nlg we obtain @xmath106a@xmath104 whereas with the lm2m2 potential the result is @xmath107a@xmath104 . this shows that the level functions produce a description with the accuracy better than @xmath108 . the results for the three - body parameter corresponding to the first excited state are @xmath109a@xmath104 and @xmath110a@xmath104 for the lg and nlg respectively . to be compared to the lm2m2 value of @xmath111a@xmath104 . as expected finite range effects are reduced in this level . the previous analysis can be extended to the four - body case . in this case a two - level structure , with energies @xmath112 and @xmath113 , is attached to each @xmath13 level @xcite . as in the three - body case , the four - body system can be studied using potential models , also in this case there is a modification of the universal ratios due to finite - range effects ( see for example refs . @xcite ) . following the previous discussion , the equations describing the four - boson spectrum can be written as @xmath114 [ eq : energyfr4 ] with @xmath115 and @xmath11 identifying the corresponding three - body branch . the four - body parameters @xmath116 are related to the energy of the level at the unitary limit , @xmath117 . it should be stressed that only the branch @xmath51 , with energies @xmath118 and @xmath119 , corresponds to true bound states . the other states , corresponding to branches with @xmath120 , are above the trimer ground state threshold and therefore are unstable bound states ( ubs ) @xcite . in the above equation we have introduced the level function @xmath121 that governs the four - body spectrum in levels @xmath122 . they can be computed using potential models using the following definition @xmath123 for @xmath51 it could be very different from the universal function @xmath124 that governs the four - body spectrum in the zero - range limit . however , as @xmath11 increases finite - range effects become negligible and @xmath121 should tend to that function . the results are shown in fig . [ fig : fig4 ] where the function @xmath121 has been calculated using the lg and nlg potentials for the ground state , @xmath125 level ( squares ) and first excited state , @xmath126 level ( triangles ) . for the sake of comparison the function corresponding to the @xmath127 level ( circles ) and calculated using the nlg is also shown . from the figure we can see that the results of both potentials for the @xmath125 and @xmath126 levels are close to each other . however there is a difference between the functions with different @xmath10 values , more pronounced for @xmath128 . eq.([eq : energyfr4 ] ) works very similar to the three - body case discussed before . the knowledge of @xmath121 makes this equation a one parameter equation . we would like to stress that @xmath121 can be computed using a lg or a nlg potential and then used to determine the spectrum of a real system around the unitary limit . as an example we discuss the spectrum of four he atoms . a very complete discussion of this system has been given in refs.@xcite using realistic potentials . the binding energies of the tetramers using the lm2m2 interaction are @xmath129mk and @xmath130mk . the dimer energy is @xmath131mk and , therefore the angles are @xmath132 and @xmath133 respectively . using the nlg @xmath121 function we estimate @xmath134k and @xmath135k in a very good agreement with the quoted values for the lm2m2 potential of @xmath136k and @xmath137k given in ref . we can conclude that the estimates obtained using the level functions are given with an accuracy well below @xmath108 . for different levels and potentials . the solid green line is an interpolation of the @xmath138 values and should represent the universal function @xmath139 . ] in fig . [ fig : fig4 ] the level function @xmath138 , calculated using the nlg potential , is shown ( circles ) . for this level finite - range effects can be neglected and , therefore , we can consider this function a good representation of the zero - range four - body universal function in the level @xmath14 . for a generic level @xmath10 we define this function @xmath140 . it does not depend on the three - body branch @xmath11 , as dsi , with the geometrical factor @xmath141 , has been already verified among these branches @xcite . in the following we study its dependence on the levels @xmath10 and its relation with the three - boson universal function @xmath6 . to this aim the level functions @xmath138 and @xmath142 , calculated using the nlg potential , are shown in fig . [ fig : fig5 ] together with the zero - range universal function @xmath6 and the level function @xmath99 . we consider @xmath99 a representation of the zero - range universal function @xmath6 and we consider @xmath143 a representation of @xmath140 . for the @xmath15 case the range of @xmath144 values in which this level results an inelastic virtual state ( ivs ) is explicitly shown . from the figure we can see that @xmath6 and @xmath140 are very close to each other around the unitary limit . as the functions approach the different thresholds differences appear . in the case of the threshold at @xmath145 we have @xmath146 , appreciably different from @xmath147 . defining @xmath148 , the two - body scattering length at which the four - boson system disappears into the four - body continuum , the first relation establishes that @xmath149 is a universal number . defining @xmath150 to be the two - body scattering length at which the trimer disappears into the three - body continuum , the second relation results @xmath151 confirming the highly accurate value given in ref . @xcite . within the zero - range theory , the set of values at which the different branches disappear into the three- and four - body continumm can be determined from the above relations using the universal ratios @xcite @xmath152 , @xmath153 , @xmath154 . the dimer - dimer threshold is defined by @xmath155 corresponding to @xmath156 . in fig . [ fig : fig5 ] we can observe that around this value @xmath157 differs from @xmath158 and @xmath6 , with these last two functions close to each other . it should be noticed that in this region @xmath158 is obtained from the energy of the ivs except for the vicinity of @xmath159 where the shallow tetramer again becomes ubs before decaying through the dimer - dimer threshold . the values of the universal functions at the critical value @xmath159 can be calculated using the universal ratios of ref . @xcite : @xmath160 , @xmath161 , @xmath162 and the relation @xmath163 , with @xmath164 , @xmath165 and @xmath166 the scattering lengths at which the level @xmath122 of the tetramer intersects the dimer - dimer threshold and at which the level @xmath11 of the trimer intersects the dimer - dimer and dimer thresholds , respectively . using these ratios in eq.([eq : energyfr4 ] ) it results @xmath167 and @xmath168 , respectively , in complete agreement with the computed values shown in fig . [ fig : fig5 ] . we can conclude that the small differences between @xmath157 and @xmath158 around the critical value are due to threshold effects . ( green solid points ) and @xmath158 ( blue solid points and violet solid points in the ivs region ) . for the sake of comparison , the three - boson universal function @xmath6 and level function @xmath169 are also shown . ] from the above discussion we propose the following zero - range equation for the @xmath4 spectrum of four equal bosons @xmath170 [ eq : energyzr4 ] with @xmath171 the binding momentum of the level @xmath172 at the unitary limit verifying the following universal ratio @xmath173 and @xmath140 the universal four - boson universal function given in fig . [ fig : fig5 ] . this equation extends the efimov radial law for three bosons to the four - boson system . in the present work we have analysed two- , three- and four - boson systems close to the unitary limit . to this aim we have solved the schrdinger equation ( or fy and ags equations ) using potential models with variable strength constructed to reproduce the minimal information given by the two - body scattering length @xmath0 and the two - body binding energy or virtual state energy @xmath1 . it has been shown that a two - parameter interaction as a gaussian can capture the main ingredients of the dynamics in this region . moreover these type of potentials define level functions independent of the range used to compute it . this property can be used to construct level functions of general validity that can be used to predict some characteristics of real systems along the particular path used to reach the unitary limit . the level functions @xmath174 and @xmath175 have two properties : when they are used in the lowest branches , @xmath90 , they absorb finite - range effects . this portion of the spectrum does not show a perfect dsi since finite - range effects are visible . so the interest here is to use the level functions to describe the dynamics of real systems close to the unitary limit . for example potentials with variable strength describe with reasonable accuracy the variation of the interatomic potential using broad feshbach resonances in ultra cold atomic traps . the second property is given by the description of the asymptotic part of the spectrum . for levels with @xmath88 the spectrum calculated with the potentials starts to show dsi and coincides with the spectrum in the zero - range limit . accordingly , the level functions for @xmath88 do not depend any more on the level number @xmath11 and on the particular potential used to compute it as well as the path selected to reach the unitary limit . they are good representations of the universal functions @xmath176 and @xmath177 . this property has been used here to propose eq.([eq : energyzr4 ] ) as an extension of the efimov radial law for four bosons . in the present work we have studied a particular path to reach the unitary limit based on a single channel potential with variable strength . other possibilities could be for example the study of coupled channel interactions as in molecular systems . in this way different level functions can be constructed allowing to a systematic study of finite - range effects . other improvements of the present work could be the study of the spectrum as the number of bosons increases . preliminary results along this line have been obtained @xcite . finally we would like to mention the recent study of efimov physics in the three - body system having spin - isospin degrees of freedom @xcite . the extension to the four - body system is under way . this work was partly supported by ministerio de economa y competitividad ( spain ) under contracts mtm2015 - 63914-p and fpa2015 - 65035-p . part of the calculations of this work were performed in the high capacity cluster for physics , funded in part by universidad complutense de madrid and in part with feder funds as a contribution to the campus of international excellence of moncloa , cei moncloa . ministerio de educacin , cultura y deporte ( spain ) for the `` jos castillejo '' fellowship in the framework of plan estatal de investigacin cientfica y tcnica y de innovacin 2013 - 2016 . 99 v. efimov , phys . lett . b * 33 * , 563 ( 1970 ) v. efimov , sov.j . * 12 * , 589 ( 1971 ) , t. kraemer et al . , nature * 440 * , 315 ( 2006 ) f. ferlaino , a. zenesini , m. berninger , b. huang , h.c . n " agerl , and r. grimm , few - body syst . * 51 * , 113 ( 2011 ) o. machtey , z. shotan , n. gross , and l. khaykovich , phys . lett . * 108 * , 210406 ( 2012 ) s. roy , m. landini , a. trenkwalder , g. semeghini , g. spagnolli , a. simoni , m. fattori , m. inguscio , and g. modugno , phys . lett . * 111 * , 053202 ( 2013 ) p. dyke , s.e . pollack , and r.g . hulet , phys . a * 88 * , 023625 ( 2013 ) p.f . bedaque , h .- w . hammer , and u. van kolck , phys . rev . lett.*82 * , 463 ( 1999 ) p. bedaque , h .- w . hammer , and u. van kolck , nucl . a*676 * , 357 ( 2000 ) e. braaten and h .- w . hammer , phys . * 428 * , 259 ( 2006 ) t. frederico , l. tomio , a. delfino , m. r. hadizadeh , and m.t . yamashita , few - body syst . * 51 * , 87 ( 2011 ) l.h . thommas , phys . rev . * 47 * , 903 ( 1935 ) e. braaten , h .- w . hammer and m. kusunoki , phys . a * 67 * , 022505 ( 2003 ) l. platter , h.w . hammer , and ulf - g . meiner , phys . rev . a*70 * , 052101 ( 2004 ) h .- w hammer and l. platter , eur . j. a*32 * , 113 ( 2007 ) j , von stecher , j.p . dincao abd c.h . greene , nat . * , 417 ( 2009 ) m. r. hadizadeh , m.t . yamashita , l. tomio , a. delfino , and t. frederico , phys . * 107 * , 135304 ( 2011 ) a. deltuva , phys . rev . a*82 * , 040701(r ) ( 2010 ) a. kievsky and m. gattobigio , phys . a * 92 * , 062715 ( 2015 ) c. chin , r. grimm , p. julienne and e. tiesinga , rev . * 82 * , 1225 ( 2010 ) r.a . aziz and m.j . slaman , j. chem . phys . * 94 * , 8047 ( 1991 ) a. kievsky , l.e . marcucci , s. rosati and m. viviani , few - body syst . * 22 * , 1 ( 1997 ) m. viviani , a. kievsky and s. rosati , phys . rev.*c71 * , 024006 ( 2005 ) m. gattobigio , a. kievsky and m. viviani , phys . a * 84 * , 052503 ( 2011 ) o. a. yakubovsky , yad . fiz . * 5 * , 1312 ( 1967 ) [ sov . phys . * 5 * , 937 ( 1967 ) ] . p. grassberger and w. sandhas , nucl . phys . * b2 * , 181 ( 1967 ) ; e. o. alt , p. grassberger , and w. sandhas , jinr report no . e4 - 6688 ( 1972 ) . a. deltuva , epl * 95 * , 43002 ( 2011 ) . m. gattobigio and a. kievsky , phys . a * 90 * , 012502 ( 2014 ) a. kievsky and m. gattobigio , phys . rev . a * 87 * , 052719 ( 2013 ) h .- w . hammer , m. gattobigio and a. kievsky , in preparation a. kievsky , n.k . timofeyuk and m. gattobigio , phys . a * 90 * , 032504 ( 2014 ) a.m. badalyan , l.p . kok , m.i . polikarpov and y.a . simonov , phys . rep . * 82 * , 31 ( 1982 ) a. o. gogolin , c. mora , and r. egger , phys . * 100 * , 140404 ( 2008 ) e. hiyama and m. kamimura , phys a * 85 * , 062505 ( 2012 ) e. hiyama and m. kamimura , phys . a * 90 * , 052514 ( 2014 ) a. deltuva . a * 85 * , 012708 ( 2012 ) a. deltuva , phys . a * 84 * , 022703 ( 2011 ) .
two- , three- , and four - boson systems are studied close to the unitary limit using potential models constructed to reproduce the minimal information given by the two - body scattering length @xmath0 and the two - body binding energy or virtual state energy @xmath1 . the particular path used to reach the unitary limit is given by varying the potential strength . in this way the energy spectrum in the three- and four - boson systems is computed . the lowest energy states show finite - range effects absorbed in the construction of level functions that can be used to study real systems . higher energy levels are free from finite - range effects , therefore the corresponding level functions tend to the zero - range universal function . using this property a zero - range equation for the four - boson system is proposed and the four - boson universal function is computed .
figure s1 shows maps of the differential resistance versus bias current and gate voltage taken at base temperature for junctions b - g . oscillations of the critical current @xmath33 in the p - doped regime are easier to observe in junctions a and e , which are significantly shorter than the other junctions . the dimension of the junctions are listed in table s1 . [ cols="^,^,^",options="header " , ] generally speaking , due to damping , environmental , and thermal effects the true critical current , @xmath33 , of a small josephson junction is inaccessible ; instead a smaller switching current @xmath85 is measured . in particular , in evaluating the dynamics of the junctions ( overdamped vs. underdamped ) it is important to consider the capacitance of the large ( @xmath86 ) bonding pads coupled via a global back gate , which could contribute up to @xmath87 to the junction capacitance . in our earlier work ( ref . ) , we fabricated @xmath88 resistors in the leads in order to isolate the bonding pads , making the junctions certainly overdamped . in this work , the junctions are not purposefully isolated from the bonding pads . therefore , for the shortest junctions and at high critical currents , we estimate that the quality factor could potentially reach up to @xmath89 , the slightly underdamped regime . ( the real @xmath90 of the junction may either be reduced due to the lead impedance , or increased due to the effectively suppressed internal dissipation of the junction at low temperatures . ) we therefore chose to verify our approximation @xmath91 experimentally , by analyzing the statistical distribution of @xmath85 . the distribution is measured by repeatedly sweeping the bias current from zero past the switching , and recording @xmath85 at every sweep @xcite . thus we obtain @xmath92 , the probability that the junction will switch from the superconducting to the normal state at bias current @xmath93 . obtaining the critical current @xmath33 via statistical methods , while more accurate , would take a prohibitively long time if performed for every junction , at every gate voltage @xmath35 and temperature @xmath41 . instead , we choose one of the shorter junctions ( e , 300 nm long ) , which would more likely demonstrate premature switching due to being underdamped . we then select several representative gate voltages and verify that the measured switching current @xmath85 is very close to the true critical current @xmath33 ( figure s2a ) . instead of the probability distribution @xmath92 , it is more informative to look at the junction escape rate @xmath94 , which may be obtained by summing over the switching histogram @xcite @xmath95 \right)\ ] ] [ equ.1 ] where @xmath96 is the current sweeping rate . the resulting calculation @xmath97 allows one to estimate the true critical current @xmath33 . as the bias current approaches the critical current @xmath98 , @xmath99 should follow the relationship : @xmath100 , with @xmath101 for switching mediated by thermal activation @xcite . an example of this dependence is shown in figure s2b . by fitting this dependence , we obtain a good estimate of the critical current @xmath33 . we find that as @xmath33 increases , so does its difference from @xmath85 . however , even for the largest presented currents ( @xmath102a ) , the measured current is suppressed by not more than @xmath103 , so we are justified in using the measured @xmath85 in place of @xmath33 in the main text . figure s3 shows the critical current @xmath47 on a semilogarithmic scale for devices e , f and g. similar to devices in the main text , @xmath53 is clearly linear in @xmath41 over more than an order of magnitude , consistent with the expectation for long ballistic junctions . knowing the energy scale @xmath2 , we can examine the universality of the critical current @xmath33 in these ballistic devices . instead of the absolute temperature scale , we use the device - independent ratio @xmath104 , which accounts for the difference in junction lengths . next , we normalize the magnitude of the critical current , thus accounting for the device width , as well as the carrier concentration . the normalization point is chosen at @xmath105 , the lowest point for which data are available for all three junctions . the normalized current plotted versus @xmath104 falls on the same curve for the four junctions and different gate voltages ( figure s5 ) , which strongly indicates that the critical current is a universal function of the dimensionless ratio @xmath104 . we observed that at high density and @xmath41 approaching zero , each of the @xmath107 transverse modes contributes a supercurrent on the order of @xmath108 . this is illustrated in figure 4 of the main paper , by showing that the ratio @xmath109 tends to a constant at high @xmath35 . however this ratio is suppressed close to charge neutrality , which we attributed to the lower contact transmission close to the dirac point , as illustrated in figure 3a inset . in figure s7 we plot the ratio of @xmath110 to account for the fact that the normal resistance @xmath40 slightly differs from @xmath111 and the supercurrent carried by each mode should be reduced as @xmath112 . here we observe that the ratio saturates and becomes constant at gate voltages much closer to the dirac point , which suggests that the reduction in @xmath113 is indeed caused by the imperfect transmission at charge neutrality . y. luh , acta phys . sin . * 21 * , 75 ( 1965 ) ; shiba , h. classical spins in superconductors . _ * 40 * , 435 ( 1968 ) ; a. i. rusinov , a. i. on the theory of gapless superconductivity in alloys containing paramagnetic impurities . . jetp _ * 29 * , 1101 ( 1969 ) .
we investigate the critical current , @xmath0 , of ballistic josephson junctions made of encapsulated graphene / boron - nitride heterostructures . we observe a crossover from the short to the long junction regimes as the length of the device increases . in long ballistic junctions , @xmath0 is found to scale as @xmath1 . the extracted energies @xmath2 are independent of the carrier density and proportional to the level spacing of the ballistic cavity , as determined from fabry - perot oscillations of the junction normal resistance . as @xmath3 the critical current of a long ( or short ) junction saturates at al level determined by the product of @xmath2 ( or @xmath4 ) and the number of the junction s transversal modes . encapsulated graphene / boron - nitride heterostructures emerged in the past year as a medium of choice for studying proximity - induced superconductivity in the ultra - clean limit @xcite . these junctions support the ballistic propagation of superconducting currents across micron - scale graphene channels , and their critical current is gate - tunable across several orders of magnitude . in these devices , a rich phenomenology arises from the interplay of superconductivity with ballistic transport @xcite , cyclotron motion @xcite , and even the quantum hall effect at high magnetic field @xcite . in a superconductor - normal metal - superconductor ( sns ) junction , single particles in the n region can not enter the superconductor and therefore experience andreev reflections at each s - n interface . this results in andreev bound states ( abs ) , which are capable of carrying superconducting current across the n region . in long ballistic junctions , the energy spectrum of the abs is quantized with a level spacing of @xmath5 , where @xmath6 is the junction length and @xmath7 the fermi velocity @xcite . the energy of abs can not exceed the superconducting gap @xmath4 , so in the short junction regime , @xmath8 , only a single abs remains . in this work we study several ballistic junctions of different length and demonstrate that the temperature dependence of the critical current dramatically differs in the long and short regimes . for long junctions , we observe an exponential scaling of the current through the junction @xmath9 , where @xmath10 @xcite . note that in graphene @xmath7 is a constant , and @xmath2 is expected to be independent of the carrier density or the mobility ( as long as the junction remains ballistic . ) for comparison , in a short junction we observe a different scaling , as expressed in eq . ( 1 ) , in excellent agreement with the theory @xcite . our graphene layers are exfoliated from kish graphite and encapsulated in hexagonal boron - nitride ( hbn ) using the `` pick - up '' method @xcite . heating beyond 250@xmath11c causes bubbles of trapped adsorbates to migrate towards the edges of the graphene mesa , effectively cleaning it . the edges of the graphene flake are exposed by etching through the hbn - graphene - hbn stack with a chf@xmath12/o@xmath13 plasma ( flow rates @xmath14 sccm ) at @xmath15pa and @xmath16w power . the etching time varies depending on the thickness of the top hbn layer . we use dc magnetron sputtering to form molybdenum - rhenium alloy contacts ( @xmath17 wt@xmath18 ) , with a measured superconducting gap @xmath19mev ( figure 1b ) . these contacts are @xmath20 nm thick and are deposited at a rate of @xmath21nm / min ( with a pressure of @xmath22mtorr and a power of @xmath23w @xcite ) . in this work we studied seven josephson junctions with lengths ranging from 200@xmath24 nm to 2000@xmath24 nm . device dimensions are listed in the supplementary information @xcite . junction a is found to be in the short regime , junctions b and e are intermediate , while junctions c , d , f , and g are in in the long regime . below we present primarily the data measured on four junctions a - d ( @xmath25 200 nm , 400 nm , 1@xmath26 m , and 2@xmath26 m ) fabricated on the same substrate . the junctions are measured in a four - terminal setup with the carrier density in graphene being controlled by a gate voltage , @xmath27 . figure 1 presents a map of the differential resistance @xmath28 , measured on junction a at @xmath29 k. the dark region of vanishing resistance indicates a supercurrent , which persists at all values of @xmath27 . as the current is swept from the negative to the positive values , the transition from the normal to the superconducting state is seen at negative bias when @xmath30 ( the retrapping current . ) the transition from the superconducting back to the normal state happens at positive bias when @xmath31 . as commonly observed in graphene josephson junctions , at low temperatures the samples exhibit hysteresis , @xmath32 @xcite , which could be attributed to either underdamped junction dynamics @xcite , or to the self - heating by the retrapping current @xcite . as discussed in the supplementary material , the second scenario is more likely for most of the range studied here . based on the measurements of the switching statistics @xcite , in the following we will use the switching current to represent the true critical current of the junction , @xmath33 . in the hole - doped regime , the reflections of ballistic charge carriers from the n - doped contact interfaces yield the quantum ( `` fabry - perot '' ) interference . a very similar oscillation pattern could be observed in the dependence of both the the normal conductance , @xmath34 , and the critical current @xmath33 on gate voltage @xmath35 ( figure 1c ) @xcite . oscillations are also observed as a function of bias voltage @xmath36 ( figure 4a inset ) @xcite . the critical current @xmath33 is observed to rapidly decrease with temperature , however the functional form of @xmath37 strongly depends on the length of the junction . figure 2 shows the evolution of @xmath37 from the short to the long regime . each panel shows data measured for several values of @xmath27 , which from here on is shown relative to the dirac point . the shortest junction ( figure 2a ) can only support a single abs ; in this regime , the current is : @xmath38 where @xmath39 is the transmission coefficient of the s - n interface and @xmath40 the normal state resistance . for a given @xmath41 , this expression should be maximized over @xmath42 to determine @xmath37 @xcite . moreover , at higher temperatures the superconducting gap will be suppressed ; we approximate the temperature dependence of the gap as @xmath43 , where @xmath44 is the gap for @xmath45 , and @xmath46 is the critical temperature @xcite . taking the complete temperature - dependent expression , we fit @xmath47 for junction a using the value @xmath44@xmath48 1.2 mev extracted from multiple andreev reflections measurements . the fit is in excellent agreement with the data ( figure 2a ) . a ) normal conductance of junctions a - d normalized by the width of the junctions , @xmath49 . even though the device lengths are different by up to a factor of @xmath50 , the three curves are very close to each other , thus proving the ballistic nature of these junctions . at positive @xmath35 , @xmath34 of all junctions is found to approach @xmath51 ( gray dashed line ) , indicating consistently high contact transparency for n - doping in these devices . inset : transmission coefficient @xmath39 of junction a. @xmath39 is calculated via two methods : comparing the normal conductance @xmath34 to the ballistic limit @xmath52 ( blue ) , and fitting the critical current @xmath33 vs. temperature @xmath41 ( red ) . both methods provide consistent results and indicate high contact transparency for n - doping . b ) energy @xmath2 extracted from the slope of @xmath53 vs. @xmath41 for junctions ( b - g ) . as expected in the long junction regime , @xmath2 depends only on device length @xmath6 and is almost density - independent through both the electron and hole doping . ] the transmission coefficient @xmath39 extracted from the fit is plotted in the inset of figure 3 as a function of the gate voltage . we can also estimate the transmission coefficient via an alternative method , by comparing the junction normal conductance @xmath34 to the ballistic limit of conductance , @xmath54 , where @xmath55 is the number of transversal modes and @xmath56 is the carrier density . @xmath39 estimated as @xmath57 is shown in blue in the inset of figure 3 . both methods provide consistent results , with @xmath39 in junction a reaching 90@xmath18 at high densities . furthermore , we find that the normal conductance of all junctions is very close to the ballistic limit . figure 3a compares the normal conductance of junctions a - d normalized by junction s width ( in fact , junctions a - c have the same widths ) . all four curves are very close to each other and approach the ballistic limit for positive gate voltages ( dashed line ) . this result indicates two important facts : a ) the contacts of all junctions are highly transparent on the n - doped side and b ) the junctions conductances do not depend on length , confirming their ballistic nature . we now return to the critical current measured in the longer junctions b - d . in figure 2(b - d ) , @xmath33 is plotted on a semilogarithmic scale and clearly shows exponential dependence at high temperatures @xmath41 ( over an order of magnitude in panels c and d ) . this is consistent with the expected long junction behavior @xmath9 @xcite and allows us to extract the energy scale @xmath2 . the temperature dependence eventually saturates at low temperatures , when @xmath58 becomes comparable to @xmath2 . figure 3b shows that for a given device @xmath59 remains roughly constant as a function of @xmath27 for both electron and hole doping , as expected in the long ballistic regime . @xmath2 is on the order of 0.05 mev for the longest device , junction d , and goes up to @xmath60 mev for junction e. while @xmath2 is consistent with the expected value of @xmath61 for junction d , it is suppressed for shorter junctions . as the devices are ballistic , the suppression of @xmath2 can not be explained by the effective lengthening of the carrier path due to diffusion . to explain the suppressed @xmath2 , we observe that the previous discussion of the long junctions neglected the coherence length @xmath62 compared to @xmath6 . taking @xmath62 into account suppresses the level spacing , which becomes @xmath63 @xcite . while the general expression for @xmath64 in the @xmath65 regime is not known , numerical simulations show that it still roughly follows the @xmath66 dependence , with @xmath2 suppressed by a factor of @xmath67 compared to the estimate that neglects @xmath62 ( figure 3b in ref . @xcite ) . in our case , @xmath68 nm , which explains the suppressed @xmath2 in the intermediate regime ( junctions b , e ) . eventually , the junction transitions to the short regime , where the exponential dependence no longer holds . we now turn to the saturation of @xmath0 in the low temperature limit : @xmath69 for a short junction , or @xmath70 for a long junction . in the long ballistic junction regime , the @xmath71 critical current is expected to be on the order of @xmath72 per transversal mode @xcite . figure 4b shows the ratio @xmath73 as a function of the gate voltage . strikingly , the curves for the four junctions are very close to each other and converge to a constant level of @xmath74 at high gate voltage , where the graphene - more interfaces are highly transparent . ( see supplementary for data on additional devices . ) similarly , the @xmath71 critical current per mode is expected to be @xmath75 in an ideal short junction @xcite . figure 4a plots the ratio @xmath76 for junction a , which indeed saturates at high gate voltage , although its value @xmath77 is significantly smaller than @xmath67 predicted by theory of ref . @xcite . previous works have observed similar deviations from theory @xcite . the mechanism for such suppression is unclear and can not be explained by environmental damping effects , nor the effect of imperfect transmission @xcite . ( note : as there are currently no graphene - specific theoretical works predicting the ratio @xmath73 in the long regime , it is unclear whether the value of @xmath78 observed in figure 4b is coincidental . ) the ratio @xmath73 is significantly reduced close to charge neutrality . this suppression most likely arises from the @xmath35 dependence of the transmission coefficient @xmath39 of the superconductor - graphene interface . we extract the contact transparency from the junction normal resistance as @xmath79 and find that while @xmath39 is close to 1 at high densities , it does get significantly suppressed close to the charge neutrality point . considering this suppression allows us to partially account for the reduced @xmath73 ratio ( see supplementary ) . in conclusion , we studied the nature of the critical current in several ballistic superconductor - graphene - superconductor junctions . we find that in the short junction regime , @xmath80 , the critical current follows eq . ( 1 ) , while in the intermediate and long junctions @xmath0 is @xmath81 . the slope of @xmath82 vs. @xmath41 dependence allows us to extract the energy scale @xmath2 , which depends on the junction length but not the gate voltage @xmath35 . while consistent for very long junctions @xmath83 , the values of @xmath2 for intermediate devices @xmath84 are smaller than those naively estimated from the junction lengths . we attribute this suppression to the finite coherence length . finally , we show that at the lowest temperature , @xmath0 saturates at a level determined by the product of @xmath44 or @xmath2 ( depending on the regime ) , and the number of transversal modes across the junction width . our observations demonstrate the universality of the critical current in several regimes relevant to most hybrid superconductor - encapsulated graphene devices . i.v.b . and m.y . acknowledge the canon foundation . c.t.k . , m.t.w . , a.s . , and g.f . were supported by aro award w911nf-16 - 1 - 0122 . f.a . acknowledges the aro under award w911nf-14 - 1 - 0349 . a.w.d . was supported by the nsf graduate research fellowship dgf1106401 . low - temperature measurements performed by g.f . were supported by the division of materials sciences and engineering , office of basic energy sciences , u.s . department of energy , under award de - sc 0002765 . this work was performed in part at the duke university shared materials instrumentation facility ( smif ) , a member of the north carolina research triangle nanotechnology network ( rtnn ) , which is supported by the national science foundation ( grant eccs-1542015 ) as part of the national nanotechnology coordinated infrastructure ( nnci ) . m.y . acknowledges financial support by grant - in - aid for scientific research on innovative areas `` science of atomic layer '' . m.y . and s.t . acknowledge support by grant - in - aid for scientific research s ( no . 26220710 ) , and grant - in - aid for scientific research a ( no . 26247050 ) . we would like to thank konstantin matveev for fruitful discussions . 99 v. e. calado , s. goswami , g. nanda , m. diez , a. r. akhmerov , k. watanabe , t. taniguchi , t. m. klapwijk , and l. m. k. vandersypen , nature nano . * 10 * , 761 - 764 ( 2015 ) . m. ben shalom , m. j. zhu , v. i. falko , a. mishchenko , a. v. kretinin , k. s. novoselov , c. r. woods , k. watanabe , t. taniguchi , a. k. geim , and j. r. prance , nature phys . * 12 * , 318 - 322 ( 2016 ) . m. t. allen , o. shtanko , i. c. fulga , j. i. j. wang , d. nurgaliev , k. watanabe , t. taniguchi , a. r. akhmerov , p. jarillo - herrero , l. s. levitov , and a. yacoby , nature phys . * 12 * , 128 - 133 ( 2016 ) . f. amet , c. t. ke , i. v. borzenets , y. wang , k. watanabe , t. taniguchi , r. s. deacon , m. yamamoto , y. bomze , s. tarucha , and finkelstein , science * 352 * ( 6288 ) , p.966 ( 2016 ) . i.o . kulik , sov . phys . jetp * 30 * , 944 ( 1970 ) . j. bardeen and j. l. johnson , phys . rev . b * 5 * , 72 ( 1972 ) . a. v. svidzinskii , _ spacially - inhomogeneous problems of theory of superconductivity _ ( nauka , moscow , 1982 ) . m. tinkham , _ introduction to superconductivity _ ( mcgraw - hill , new york , 1996 ) . a. a. golubov , m. yu . kupriyanov , and e. iliichev , rev . mod . phys . * 76 * , 411 ( 2004 ) . a. v. svidzinsky , t. n. antsygina , and e. n. bratus , sov . phys . jetp * 3 * , 860 ( 1972 ) . a. v. svidzinsky , t. n. antsygina , and e. n. bratus , j. low temp . phys . * 10 * , 131 - 136 ( 1973 ) . c. w. j. beenakker , _ transport phenomena in mesoscopic systems _ pp.235 - 253 ( springer , berlin , 1992 ) . c. w. j. beenakker and h. van houten , phys . rev . lett . * 66 * , 3056 ( 1991 ) . g .- h . lee , s. kim , s .- h . jhi , and h .- j . lee , nature commun . * 6 * , 6181 ( 2015 ) . l. wang , i. meric , p. y. huang , q. gao , y. gao , h. tran , t. taniguchi , k. watanabe , l. m. campos , d. a. muller , j. guo , p. kim , j. hone , k. l. shepard , and c. r. dean , science * 342 * , 614 - 617 ( 2014 ) . see supplementary information h. b. heersche , p. jarillo - herrero , j. b. oostinga , l. m. k. vandersypen , and a. f. morpurgo , nature * 446 * , 56 ( 2007 ) . x. du , i. skachko , and e. y. andrei , phys . rev . b * 77 * , 184507 ( 2007 ) . c. ojeda - aristizabal , m. ferrier , s. guron , and h. bouchiat , phys . rev . b * 79 * , 165436 ( 2009 ) . i. v. borzenets , u. c. coskun , s. j. jones , and g. finkelstein , phys . rev . lett . * 107 * , 137005 ( 2011 ) . i. v. borzenets , u. c. coskun , s. j. jones , and g. finkelstein , ieee trans . . supercond . * 22 * , 1800104 ( 2012 ) . i.v . borzenets , u.c . coskun , h.t . mebrahtu , yu.v . bomze , a.i . smirnov , and g. finkelstein , phys . rev . lett . * 111 * , 027001 ( 2013 ) . h. courtois , m. meschke , j. t. peltonen , and j. p. pekola , phys . rev . lett . * 101 * , 067002 ( 2008 ) . u.c . coskun , m. brenner , t. hymel , v. vakaryuk , a. levchenko , and a. bezryadin , phys . rev . lett . * 108 * , 097003 ( 2012 ) . g .- h . lee , d. jeong , j.h . choi , y .- j . doh , and h .- j . lee , phys . rev . lett . * 107 * , 146605 ( 2011 ) . c. t. ke , i. v. borzenets , a. w. draelos , f. amet , yu . bomze , g. jones , m. craciun , s. russo , m. yamamoto , s. tarucha and g. finkelstein , nano lett . * 16 * ( 8) , 4788 ( 2016 ) . a. f. young , and p. kim , nature phys . * 5 * , 222 - 226 ( 2009 ) . p. rickhaus , r. maurand , m .- h . liu , m. weiss , k. richter , and c. schnenberger , nature commun . * 4 * , 2342 ( 2013 ) . i. hagymasi , a. kormanyos , and j. cserti , phys . rev . b * 82 * , 134516 ( 2010 ) . r. dougherty , j. d. kimel , _ superconductivity revisited _ , ( crc press , london 2013 ) . p. f. bagwell , phys . rev . b * 46 * , 12573 - 12586 ( 1992 ) . c. ishii , prog . theor . phys . * 44 * , 1525 ( 1970 ) . m. titov and c. w. j. beenakker , phys . rev . b * 74 * , 041401(r ) ( 2006 ) . * supplementary to : ballistic graphene josephson junctions from the short to the long regime *
in its native state a protein fluctuates around a configuration corresponding to the absolute minimum of its energy landscape . however , realistic energy landscapes are characterized by many competing minima . protein folding can therefore be viewed as the process of selecting the right minimum among many others . since proteins perform their function at finite temperatures , the entropy of their native states concurs in determining their stability even when folding is driven mainly by an enthalpic bias . when only one basin of the energy landscape is significantly visited , as it happens for folded proteins , entropy is mainly contributed by fluctuations around the minimum it is all _ vibrational _ entropy . native states are highly compact structures , characterized by the presence of secondary structure motifs , such as @xmath0-helices and @xmath1-sheets . in turn , the vibrational properties of a heterogeneous system are typically influenced by the presence of simmetries and by its degree of modularity . it is therefore interesting to investigate whether a correlation exists between the presence of @xmath0 and @xmath1 structures and vibrational entropy . previous work has shown that both @xmath0-helices and @xmath1-sheets are characterized by a larger flexibility than random - coil conformations @xcite . however , a systematic study quantifying the impact of secondary structure content on the vibrational properties of native protein conformations is lacking . it is the aim of this paper to introduce the simplest coarse - grained model able to quantify the correlation between vibrational dynamics of proteins and their structural organization at the secondary level . a sensible starting point for such analysis is represented by the class of coarse - grained network models , that map a given protein structure on a network of point - like aminoacids interacting trough hookean springs . in most implementations aminoacids are taken to sit at the corresponding c@xmath2 sites and are assigned the same average mass of 120 da . in this framework , a given native structure specifies by construction the topologies of inter - particles interactions , that is the networks of connectivities and equilibrium bonds . the patterns of low - frequency collective modes are uniquely dictated by the topology of the network of physical interactions that characterizes the native configuration @xcite . in this sense , network models might be considered as the simplest tools to describe the low - frequency regime of protein dynamics . however , due to the coarse - graining element , such schemes are more questionable in the high - frequency part of the vibrational spectrum . this makes them unsuitable to compute vibrational entropies , which include nonlinear contributions from all modes , as it shows from its very definition @xmath3 where the last passage follows in the classical limit @xmath4 @xmath5 . as a consequence , one needs to incorporate further structural and dynamical details within the network description besides the sheer topology of connections in order to correctly compute vibrational entropies . more precisely , we are interested in capturing the structural and dynamical elements that characterize fluctuations at the scale of secondary structures . we identify three crucial features that must be taken into account . * a realistic hierarchy of force constants , reproducing the different strengths of bonds featured in a protein molecule . these comprise covalent bonds ( such as the peptide bond along the chain ) , and in particular the interactions stabilizing secondary structures . i.e. hydrogen bonds , besides long - range forces such as * the degrees of freedom corresponding to side chains , which play a fundamental role in the spatial organization of secondary motifs . hydrophobic or screened electrostatic interactions . * the appropriate aminoacid masses , correctly reproducing the true protein sequence . in fact , as shown in fig . [ f : masses ] , residues of different mass show different @xmath0 and @xmath1 propensities . in particular light residues tend to be less represented in secondary structure motifs . as we will show later , this analysis highlights that @xmath1-rich ( and @xmath0-poor ) native conformations tend to have a higher vibrational entropy per residue regardless of protein size and shape . this is particularly interesting with respect to protein aggregation since protein aggregates are know to be rich in @xmath1-structures . in most cases , they share the same @xmath1-spine architecture characteristic of amyloid fibrils : an assembly of beta sheets perpendicular to the fibril axis . moreover , aggregation is a rather common phenomenon for all sorts of polypeptide chains regardless of their sequence @xcite . this suggests that the phenomenon should be governed by rather general laws , uniquely related to the dynamical properties of poly - peptide chains in solution . in particular , in view of the rather slow time - scales characteristic of aggregation and in view of our results on the correlation between vibrational entropy and secondary structure , it is tempting to postulate a mechanism of thermodynamic origin , that would favor the growth of structures rich in @xmath1-content under quite generic conditions . the paper is organized as follows . in the methods section we describe the database of native structures and the models used for our analysis . in the discussion section we discuss the features of vibrational spectra and the emergence of the observed correlation between vibrational entropy and secondary structure . finally , we comment on the biological relevance of our results , particularly for what concerns protein aggregation phenomena . it is our aim to conduct the most general analysis of the interplay between vibrational entropy and the content of secondary structure in a large database of protein structures . of course , we are bound to avoid repeatedly taking into account structures corresponding to homologous proteins . we therefore chose the pdbselect database , which was explicitly built to gather the largest number of protein scaffolds by keeping the structural homology between any two structures lower than 30 % @xcite . in order to illustrate the amount of structural diversity displayed by proteins from the pdbselect database , we show in fig . [ f : alpha - beta ] both the @xmath0-helix and @xmath1-sheet content of each structure as computed by means of the dssp algorithm introduced by kabsch and sanders @xcite . we see that the two measures roughly anti - correlate , their sum being gaussian - distributed around 50 % with a standard deviation of about 15% . as a further piece of information , we report in the inset of fig . [ f : alpha - beta ] the statistical distribution of chain lengths for all sequences contained in the database , which shows to decay exponentially . additionally , we found no appreciable correlation between @xmath0 or @xmath1-content and other structural indicators such as chain length or surface accessible area . on the contrary , although not surprisingly , the sum of @xmath0 and @xmath1-content correlates positively with the number of hydrogen bonds ( hb ) , which is indeed the principal physical interaction stabilizing both kind of motifs . versus @xmath1-content in the pdbselect database the inset shows the histogram of chain lengths in the database ( lin - log scale).,width=12 ] our starting point is a widely employed coarse - grained scheme , the anisotropic network model ( anm ) . originally proposed by tirion @xcite at the all - atom level , such model has been successively reconsidered by bahar and co - workers in the c@xmath2 approximation @xcite , with results in fair agreement with experiment . in the spirit of coarse - grained network models , aminoacids are modelled as spherical beads linked by hookean springs if they are close enough in the native structure , as specified through a pre - assigned cutoff distance . both anm and its scalar analogue , the gaussian network model ( gnm ) @xcite , proved very efficient in describing the low - frequency part of protein spectra , which is dominated by large - scale , collective motions of entire domains or other structural sub - units , as measured from x - ray crystallography or electron microscopy @xcite , let @xmath6 be the number of residues of a given structure and let @xmath7 denote the distance between the c@xmath2s of the @xmath8-th and @xmath9-th aminoacids in the pdb structure . then , the network of residue - residue interactions can be constructed by introducing a distance cutoff @xmath10 , and the corresponding _ connectivity _ matrix @xmath11 accordingly , the total potential energy of the c@xmath12 anm is a sum of harmonic potentials , @xmath13 here @xmath14 measures the instantaneous elongation of the @xmath15 bond and @xmath16 is the strength of all springs connecting interacting pairs . the anm model is known to describe accurately long - wavelength fluctuations , such as the concerted motions of subunits or entire domains , whereas some doubts can be cast on its predictive accuracy concerning aminoacid motions at a scale comparable to the characteristic dimension of a residue . however , we are interested in computing vibrational entropies , which depend on the whole spectrum . importantly , the contributions of high and low frequencies to the entropy are impossible to disentangle in an elementary manner , due to its strong nonlinearity . hence , in order to increase the spectral reliability of the c@xmath12 scheme and make it more accurate in the high - frequency domain , we introduce three additional features within the standard anm model . 1 . in the basic anm protocol , all springs share the same strength , in spite of the wide differences among the real forces governing residue - residue interactions in a real protein . hence , we introduce a _ hierarchy _ of spring constants , aimed at reproducing the strength of the most important interactions , that is covalent , hydrogen - bond and van der waals ( vw ) bonds . consequently , we modify the potential energy function ( [ e : anmpot ] ) as follows @xmath17 where @xmath18 , @xmath19 and @xmath20 are the connectivity matrices of the three distinct sub - networks comprising all aminoacids interacting via peptide bonds along the main chain , through hbs and vw interactions , respectively . the quantities @xmath0 , @xmath1 and @xmath21 are the corresponding spring constants , whose magnitude is taken by construction to span two orders of magnitude . we coin this scheme the hierarchical network model ( hnm ) . + spring constants customarily used to model covalent bonding are in the range @xmath22 kcal / mol / @xmath23 . here we chose @xmath24 kcal / mol / . a second order expansion of the lennard - jones potentials used to describe hb in different all - atom force fields gives values of @xmath1 in the range @xmath25 kcal / mol / @xmath23 . we tried values of @xmath1 in the range @xmath26 @xmath0 with no appreciable difference in the distinctive features of the vibrational spectra . finally , taking into account that the young modulus of covalent solids is about @xmath27 times that of a van der waals solid , we assume @xmath28 . coarse - graining protein structure at the c@xmath12 level , one neglects all degrees of freedom associated with aminoacid side chains . however , these are known to be subject to different positional constraints in secondary structure motifs such as parallel and anti - parallel @xmath1-sheets and @xmath0-helices . moreover , the entropic contribution of side chains has been recently proved fundamental in determining the free energy of protein native states @xcite hence , we also consider a variant of the hnm model where we model side chains as additional beads , whose equilibrium position we fix at the side chain center of mass in the native structure . furthermore , we place a covalent bond between each bead - like side chain and its corresponding @xmath0 carbon . in particular , given their small size , we neglect side chains of glycines , which will be only represented by their @xmath0 carbon bead . we remark that this approach is utterly similar to many other unified residue models @xcite . usually , in the framework of protein network models , one assigns to each @xmath0 carbon bead the average aminoacid mass of 120 da . in the present study , we adopt the following convention : we split the mass of each aminoacid into a constant contribution of 56 da , that represents the atoms lying along the peptide bond , and assign a variable mass ranging from 35 da ( for alanine ) to 150 da ( for tryptophan ) to the bead representing the side chain . in the following , we will present results for an _ hybrid _ hnm model , whereby the above three features shall be switched on and off in the calculation of vibrational entropies . our aim is to understand how the corresponding three physical properties affect vibrational spectra in general and vibrational entropy in particular . it is our aim to explore the correlation of vibrational entropy with secondary structure content across the whole database . more precisely , the _ fil rouge _ of our analysis is to check whether entropic contributions might favor the formation of @xmath1-sheets at high temperatures , thus in turn disfavoring @xmath0-helices . in view of the positive correlation existing between the @xmath0 and @xmath1 contents , ( see again fig . [ f : alpha - beta ] ) we shall consider in the following the correlation of entropy with a global structural indicator defined as ( @xmath1-content @xmath29 @xmath0-content ) , which we shall refer to as @xmath1-preference . positive correlation of this quantity with entropy leads to a negative correlation with free energy at high temperatures , that will in turn favor the formation of @xmath1-like structures . conversely , a negative correlation would signal a preference towards @xmath0 helices . importantly , in order to compare proteins of different lengths , we shall always consider _ intensive _ observables , namely entropies per residue . a first analysis of the vibrational entropy of the pdbselect structures performed by means of a standard anm model yields somewhat unintuitive results . the intensive vibrational entropy @xmath30 does not correlate with the @xmath0-helical content ( correlation coefficient 0.14 ) and anti - correlates very weakly with the @xmath1-sheet content ( correlation coefficient 0.34 ) . this result , implying that @xmath1-sheet are mildly stiffer than @xmath0-helices , is manifestly at odds with the well documented rigidity of @xmath0-helices . we conclude that the anm model , although reliable at the low - frequency end of the vibrational spectra does not include a sufficient amount of detail to realistically reproduce full vibrational entropies . as a first trial towards increasing the level of detail , we consider the bare hnm model , that is we introduce a three - level hierarchy of spring strengths in the framework of the standard anm scheme as from point ( i ) above . moreover , we shall switch on the different force constants one at a time first , in order to investigate the effects of the different physical interactions upon vibrational entropies . a first observation is that a roughly ten - fold separation in the intensity of force strengths is enough to induce quasi - additivity of the vibrational spectrum . in fig . [ f : densities ] , we show how the vibrational spectrum of sequences ` 1kfm ` ( a purely @xmath0-helical protein ) , ` 1qj9 ` ( a @xmath1-barrel ) and ` 1lsx ` ( 36% @xmath0 and 36% @xmath1 ) change upon sequentially switching on the three different interactions . ) , hydrogen bonds ( middle row , @xmath31 ) and van der waals ( bottom row ) interactions for an @xmath0-helical ( panels a , b , c @xmath29 pdb code ` 1kfm ` ) , a @xmath1-barrel ( panels d , e , f @xmath29 pdb code ` 1qj9 ` ) and a structure with equal @xmath0 and @xmath1 content ( panels g , h , i @xmath29 pdb code ` 1lsx ` ) . parameters are : @xmath32 , @xmath33 kcal / mol / @xmath23 , @xmath34 kcal / mol / @xmath23 and @xmath35 kcal / mol / @xmath23.,width=12 ] as a general remark , we observe that , for all types of structure , the bare covalent chain has only @xmath6 modes with non - zero frequency . more in detail , we see that the structure with dominant @xmath0-helical character has a very distinctive peak at high frequency . moreover , such feature shows to be robust with respect to the nature of residue - residue interactions . a peak is also seen to emerge in the @xmath1 structure as soon as the connectivity matrix starts having its off - diagonal regions populated due to hbs and vw interactions . more quantitatively , we found that the high - frequency part of the spectra across all the pdbselect could be well reproduced by a linear combination of two normalized step functions @xmath36 and @xmath37 , with supports lying , respectively , in the @xmath38 $ ] @xmath39 and @xmath40 $ ] @xmath39 intervals as shown in fig . [ f : fitting ] . the two resulting coefficients @xmath41 and @xmath42 sum up to unity and might be regarded as a spectral estimate of the actual @xmath0 and @xmath1 contents . at a closer inspection , it turns out that only @xmath41 is a good approximation of the actual @xmath0-content of the protein ( see fig . [ f : formfactor - alpha ] ) , while @xmath42 is observed to be large not only when the protein has an high beta content but also when it has no pronounced secondary structural features at all . in other words , while the @xmath0-peak is a clear distinctive feature of @xmath0-helices , @xmath1-sheets are rather spectrally indistinguishable from unstructured regions ( see the appendix for a quantitative explication of the @xmath0-peak ) . -content by fitting the high - frequency portion of the spectrum in a typical case ( mixed @xmath43 structure ` 1lsx ` ) . the two representative functions @xmath36 ( @xmath0-peak ) and @xmath37 ( @xmath1-band ) are combined linearly in the step function @xmath44 , which is least - square fitted to the high - frequency portion of the spectrum . parameters as in fig . [ f : densities].,width=12 ] -content as determined from the spectra , @xmath41 , and the actual @xmath0-content as measured through the dssp protocol for all structures from the pdbselect . parameters as in fig . [ f : densities].,width=12 ] if we now turn the hbs on , vibrational spectra acquire a varying amount of new vibrational modes ( between 5% and 30% ) , on average 18% of the total number of degrees of freedom on the entire pdb select . perhaps not surprisingly , the quantity of such new modes turns out to be linearly correlated with the number of hbs ( see fig . [ f : percentage - newmodes ] ) . furthermore , we see that the appearance of the new hb part of the vibrational spectrum only slightly modifies the spectral density at higher frequencies ( compare panels b and c , e and f or h and i in fig . [ f : densities ] ) . finally , adding the lennard - jones contribution , all the remaining modes are filled , with the exception of the expected six zero - frequency modes corresponding to rigid translations and rotations . once again , due to the lower intensity of their driving interaction , the new lj modes occupy the low - frequency end of the spectrum and do not modify appreciably the spectral density at higher frequencies . .,width=12 ] the marked differences in the spectra of @xmath0 and @xmath1 proteins described so far through the hnm c@xmath2 model do not show up in the same clear - cut fashion if one examines how the vibrational entropy depends on the type of secondary structure of the proteins : the intensive entropy shows no correlation at all with @xmath1-preference ( data not shown ) . at a deeper thought , this is scarcely surprising , since the only detectable spectral differences between proteins of different secondary organization have already seen to be located at high frequencies , which in turn results in a negligible contribution to vibrational entropy . what is more , that part of the spectrum is also the most sensitive to the structural details of the system and consequently the least reliably captured by a coarse - grained model . in order to test the solidity of our conjectures , we must then increase the level of descriptive accuracy of the model . we have devised two ways to accomplish this , by adding additional degrees of freedom representing the side chains or adding the full sequence of aminoacid masses . we found that both additional features independently produce the same effect , namely destroy the clear @xmath0-peak versus @xmath1-band picture illustrated above . however , the spectral additivity is preserved and , all the more significantly , a good correlation between intensive vibrational entropy and @xmath1-preference appears , as can be clearly seen in fig . [ f : correlationssidechain ] where the two observables are scatter - plotted for the full , weighted hnm model with side chains . with @xmath1-preference in the weighted hnm model with side chains . parameters are : @xmath32 , @xmath33 kcal / mol / @xmath23 , @xmath34 kcal / mol / @xmath23 and @xmath35 kcal / mol / @xmath23.,width=12 ] a systematic analysis reveals that such correlation is shaped by a complex interplay between the features that distinguish our model from the basic anm scheme . in table [ t : models ] the arousal of the correlation between entropy and @xmath1-preference is summarized by reporting both the correlation coefficient and the slope of the first order least - square fit to the data . the two best correlations achieved are reported in bold . the best correlation is obtained when all levels of detail are present . in order to quantify their different contribution , it is useful to compute the average correlation coefficients over the four instances where a single feature is always present . for example , the average correlation coefficient is only 0.27 when side - chain coordinates are included ( last four rows of table [ t : models ] ) and 0.29 in the four cases where true masses are taken into account ( rows 2,4,6,8 ) . on the contrary , the average correlation coefficient of 0.39 obtained by sticking to realistic interaction strengths points to a crucial role of the protein dynamical heterogeneity in reproducing vibrational modes associated with different secondary structure motifs . overall , the same conclusions can be drawn by calculating the correlation drops resulting from the individual elimination of single features from the complete model ( last row in table [ t : models ] ) . in particular the importance of force hierarchy is confirmed ( a correlation drop of about 0.5 from rows 8 to 6 ) . moreover , this analysis reveals that considering equal masses also leads to the same correlation drop ( rows 8 to 7 ) . on the contrary , the elimination of side chains appears less traumatic with a limited loss of correlation ( rows 8 to 4 ) . in summary , concerning the interplay between secondary organization and vibrational properties , the three levels of detail introduced can be ranked as follows : the hierarchy of physical interactions proves to be the most important feature , followed by mass heterogeneity and , finally , inclusion of side - chain ( coarse - grained ) degrees - of - freedom . .summary of the effect of physical features of the hnm model on the correlation between vibrational entropy and @xmath1-preference : @xmath45 feature on , @xmath29 feature off . the first row corresponds to the bare anm model . [ cols="^,^,^ , > , > " , ] [ t : models ] this correlation reflects a tendency of @xmath1-rich architectures to host more low - frequency modes . this mechanism can better illustrated by switching off the weakest force constant in the complete model , that is shut down lj bonds . in this case the @xmath46 vibrational modes with non - zero frequency describe the oscillations of the network of interactions determined by hbs and covalent bonds alone , and represent all the oscillations with wavelength shorter than or equal to the typical size of secondary structure motifs . the ratio @xmath47 , with @xmath48 the number of non - zero frequency vibrational modes in the full mode might be regarded as the fraction of modes determined by the hb and covalent interactions . this assumption appears reasonable in view of the spectral quasi - additivity discussed above . more precisely , we expect that zero - frequency modes will be filled when recovering lj interactions , without significantly affecting the high - frequency portion of the vibrational spectrum . [ f : fastmodes ] shows that @xmath49 correlates negatively with the @xmath1-preference . thus , protein structures with large @xmath1-preference host fewer high - frequency modes and accordingly more low - frequency ones . this , in turn , leads to larger vibrational entropies for @xmath1-rich structures . more explicitly , we have shown that eliminating all contacts but those contributed by secondary structure motifs a clear separation of @xmath0 and @xmath1-like vibrational spectra emerges . this is further illustrated by switching off both van der waals interactions and hbs ( data not shown ) . in this case we no longer observe a correlation between @xmath49 and @xmath1-preference . thus , chain topology alone does not introduce any clear spectral signature of secondary structure organization . , the fraction of modes determined by covalent and hydrogen bonds , and @xmath1-preference over the pdbselect database using the full hnm model with side chains and proper mass sequences . the correlation coefficient is 0.44 . parameters are : @xmath32 , @xmath33 kcal / mol / @xmath23 and @xmath34 kcal / mol / @xmath23 and @xmath35 kcal / mol / @xmath23.,width=12 ] in this work we have scrutinized a wide database of non - homologous protein structures , with the aim of assessing to what extent vibrational entropy is a sensitive measure of organization at the secondary structure level . to this end , we have introduced a minimally featured coarse - grained model , that we have coined hierarchical network model ( hnm ) . at variance with current , stat - of - the - art schemes , our recipe ( i ) includes a hierarchy of physical interactions separating the strongest , short - range forces from the weaker , long - range ones , ( ii ) expands the number of degrees of freedom with the inclusion of side - chain coordinates and ( iii ) accounts for the appropriate aminoacid masses . a thorough analysis of all the structures in the data set has allowed us to unveil and quantify the correlation existing between the vibrational entropy of native folds and their specific secondary structure arrangement . more precisely , we found that _ all _ the three above - listed features of the hnm are essential in order to spotlight and quantify such correlation . remarkably , the statistical significance associated with the mixed spectral - structural signature attains its maximum value only for the full - featured scheme . of remarkable interest is the special role played by the requirement of force heterogeneity . in fact , we have proved that the presence of a realistic hierarchy of force constants accounts for the largest contribution to the observed correlation . as a final observation , it is interesting to discuss our results in the perspective of the widespread , yet still much debated phenomenon of protein aggregation . more precisely , concerning what physical effect could be held responsible for the overwhelming @xmath1 preference of mature peptidic aggregates @xcite . the lack of clear signatures of the aggregation propensity at the sequence level @xcite suggests that the phenomenon should be governed by rather general laws , uniquely related to the dynamical properties of poly - peptide chains in solution . in particular , in view of the rather slow time - scales characteristic of aggregation , it is reasonable to postulate a mechanism of thermodynamic origin , that would favor the growth of structures rich in @xmath1-content under quite generic conditions . of course , forces of enthalpic origin represent the strongest interactions along typical aggregation pathways . however , there is no reason a - priori for the strongest forces to also encode the observed bias toward @xmath1-rich structures . in fact , a composite structure developing from the aggregation of poly - peptides would likely show a tendency toward non - specific organization at the secondary structure level , thus realizing _ minimal frustration _ architectures @xcite . in view of the above facts , it is tempting to rationalize the aggregation pathway as _ driven _ by strong , non - specific forces but _ biased _ toward high content of @xmath1-type motifs by contributions to the total thermodynamic force that are weaker in magnitude but favor @xmath1-rich architectures due to their higher density of vibrational modes at low frequencies . hence , based on the results reported in this paper , one may speculate that the existing bias toward @xmath1-rich mature aggregation products be provided by a free - energy gain in vibrational entropy . following this speculation , the bulk of the free - energy changes occurring during aggregation would be mainly determined by increase of residue - residue contacts and decrease of solvent - exposed surface , while even tiny differences in vibrational entropy could be able to steer an aggregating system toward a @xmath1-rich configuration . in order to rationalize the emergence of the @xmath0-peak in the hnm - c@xmath2 model , we diagonalize the hessian matrix around a class of regular chain configurations that include both @xmath0-helices and beta sheets : those characterized by fixed values of both the bond and dihedral angle . the configurations of a linear polymer composed of @xmath6 consecutive residues can be described by a vector @xmath50 or , alternatively , once @xmath51 is known , by means of the three vectors @xmath52 with @xmath53 . let us now define a configuration of known dihedral angle @xmath54 according to the following three rules , @xmath55 @xmath56 ( bond angle must be constant ) and @xmath57 ( dihedral angle must be constant ) . once @xmath58 , @xmath59 and @xmath60 are known , @xmath61 can be found solving the system of two linear equations [ bond ] and [ dihedral ] and than by imposing the constraint [ r ] . in the simplest case , @xmath62 , the shortest chain on which a dihedral angle can be defined , one can easily show that the eigenvalues of the hessian of the potential @xmath63 are nine null ones plus @xmath64 in other words they only depend on the bond angle and they tend to peak at @xmath65 for @xmath66 . this last feature nicely explains the @xmath0-peak and the @xmath1-band , being the average c@xmath2-c@xmath2 bond angle in an @xmath0-helix just slightly lower than @xmath67 , while it is around @xmath68 for a @xmath1-sheet .
in this paper we analyze the vibrational spectra of a large ensemble of non - homologous protein structures by means of a novel tool , that we coin the hierarchical network model ( hnm ) . our coarse - grained scheme accounts for the intrinsic heterogeneity of force constants displayed by protein arrangements and also incorporates side - chain degrees of freedom . our analysis shows that vibrational entropy per unit residue correlates with the content of secondary structure . furthermore , we assess the individual contribution to vibrational entropy of the novel features of our scheme as compared with the predictions of state - of - the - art network models . this analysis highlights the importance of properly accounting for the intrinsic hierarchy in force strengths typical of the different atomic bonds that build up and stabilize protein scaffolds . finally , we discuss possible implications of our findings in the context of protein aggregation phenomena . * author summary * the intricate structure / dynamics / function relations displayed by proteins are at the core of life itself . yet , a thorough and general understanding of such interplay still remains a formidable task , at the frontier of physics and biology . proteins perform their functions while fluctuating at room temperature about their native folds . as a consequence , the entropic contribution to their free energy landscapes , i.e. vibrational entropy , constitutes a crucial element to decipher the dynamical bases of protein functions . in this study , we examine a whole database of highly - non - homologous protein structures in the effort of rationalizing the entropic contributions of the distinct secondary structure motifs in protein scaffolds . with the help of a novel coarse - grained model , we measure a significant correlation between secondary structure content and vibrational entropy , thus shedding light into the structural roots of protein flexibility . finally , we discuss our findings in the context of the unsolved problem of protein aggregation .
coulomb drag between current carrying systems has been studied in numerous papers both experimentally and theoretically@xcite . the earliest experiments were done in two dimensional systems . in this configurations the electrons are combined to two - dimensional layers . in one such layer a current is driven and another conducting layer is brought to within close proximity of the first . the coulomb interaction between the electrons in these two layers causes a transfer of momentum and the second layer acquires a current . this current is referred to as the drag current . recently , the phenomenon of coulomb drag in one dimensional systems has attracted much attention , particularly with regards to nanowires and nanotechnology @xcite . the basic description of the one dimensional case is identical to that in two dimensions . in all of these systems the time averaged charge density of the systems are translationally invariant . the coupling of the two systems is due to charge fluctuations in the two systems . it is known however that the long range character of the coulomb interaction can lead to the formation of a wigner crystal in which the electrons become localized and form a periodic lattice . numerical simulations on one dimensional systems indicate that an arbitrarily weak long range interaction will lead to the formation of a wigner crystal@xcite . this long range nature of the coulomb force should be apparent for low electron densities . in this paper we consider the drag between two such wigner crystals . it is evident that this system has a non - uniform charge distributions and we will show that this in principle can lead to a much higher value of the drag current . it is the aim of this section to analyze from a classical perspective some aspects of current drag that we may expect in a quantum mechanical treatment of the wigner crystal . the wigner crystal is as a system where the electrons are effectively localized , so it is reasonable to expect that some aspects of the current drag between wigner crystals may be exhibited in a classical system . here , we discuss the current drag between two one dimensional wires consisting of classical particles interacting via an unscreened coulomb potential . the two systems are close enough that they are coupled by the coulomb interaction between the wires . this system is illustrated in figure ( [ fig:2wires ] ) . drives system one . @xmath3 is the interwire interaction and @xmath4 is the distance of separation . ] to get an idea of what to expect from such a system we first consider a `` toy model '' consisting of two lines of equally spaced particles interacting via a periodic potential and moving under a viscous drag force . the main difference between this system and that in figure ( [ fig:2wires ] ) is that the spacing between particles is fixed so that the internal dynamics of the system are ignored . the description of such a system has been considered in an analysis by cladis , parks and daniels@xcite in the context of the vortex dynamics in type ii superconductors . in this analysis , two rigid chains of equally spaced particles are taken in the presence of a viscous drag force @xmath5 . since each chain is rigid , its position can be denoted by the position of one of its particles . call @xmath6 the position of chain @xmath7 and @xmath8 the position of chain @xmath9 . the interaction between the two chains is then periodic in @xmath10 and modeled by @xmath11$]@xcite where @xmath12 is the lattice spacing . in addition , the first chain is driven by a constant force @xmath13 . in steady state the equations of motion of the two chains are given by @xmath14 + f\ ] ] @xmath15\ ] ] setting @xmath16 , results in the following equation for the difference between the velocities of the two chains . @xmath17 with solution @xmath18 where @xmath19 is the time for the two chains to slip one complete cycle ( distance @xmath20 ) . carrying out the integral results in @xmath21 this simple analysis shows the interesting result that for suitable conditions of the parameters @xmath13 and @xmath22 , the two chains will be locked together with the drag current equal to the drive current . we now wish to investigate the behavior of this classical system when internal dynamics are included . we take therefore as a model of the dynamics of these classical particles the following fokker - planck equation @xmath23 here , @xmath24 labels the positions of @xmath25 particles in wire @xmath7 , @xmath26 labels the positions of particles in wire @xmath9 , @xmath27 is the diffusion constant and @xmath28 is the temperature . @xmath29 is then the probability of having the @xmath25 particles in layer @xmath7 at positions @xmath30 and the @xmath25 particles in layer @xmath9 at position @xmath31 at time @xmath32 . the forces present are @xmath33 in wire @xmath7 and @xmath34 in wire @xmath9 . the interaction @xmath35 is taken to be the unscreened coulomb interaction between the wires , @xmath36 is the unscreened coulomb interaction within each wire and @xmath13 is a constant force driving the particles in wire @xmath7 . the basic problem is to determine the drag current produced in wire @xmath9 due to the moving charges in wire @xmath7 . in the numerical analysis of this problem we performed monte carlo simulations on discrete systems . the conducting systems are partitioned into a lattice with the particles occupying positions at the lattice points . the dynamics of the simulation are determined by equation ( [ eq : fp ] ) which is approximated numerically as follows . for each particle , one of two possible random directions is chosen . if the particle lowers its energy by moving one lattice site in that direction , then the move is accepted . otherwise the move is accepted with probability @xmath37 , where @xmath38 and @xmath39 is the change in the particles energy upon making this move . this process is completed iteratively for each particle until a steady state condition is reached . simulations were performed to determine the dependence of the drag current on separation between the two systems . the results are shown in figure ( [ fig:1dplots ] ) the plot of drag current vs. separation shows the fact that the sum of the drive and drag currents is a constant , @xmath40 as can be seen by averaging the right hand side of equation ( [ eq : fp ] ) , ( @xmath41 ) . it also shows that the drag current is significant only when the distance of separation is on the order of the interparticle spacing or smaller . the main feature of interest in this plot , however , is that the drag current , @xmath42 , is essentially equal in magnitude to the drive current , @xmath0 , at small interwire seperations . as the separation between the wires is increased the drag current shows a smooth decrease in magnitude . the main difference between this plot and the toy model ( where internal dynamics are ignored ) is that there does not appear to be an abrupt locking transition . however at sufficiently large values of the interaction parameter , @xmath22 the two chains are essentially locked together with the drag current equal in magnitude to the drive current . we will now examine to what extent this classical effect is exhibited in a quantum mechanical system . in this section we calculate the non - dissipative drag between two perfect wigner crystal rings at zero temperature . the basics of non - dissipative drag effects has been considered in previous papers @xcite . the basic system of interest here consists of two metallic rings one of which is threaded by an ahranov - bohm flux , @xmath43 . this flux shifts the ground state , @xmath44 , of the ring to a current carrying state which is calculated according to @xmath45 if another ring is brought close to the first one , then the coulomb interaction will cause a drag current to be produced in the second ring . since this is the ground state of the two ring system this current is non - dissipative and persists as long as the magnetic flux is present in the first ring . in these works the two rings have uniform charge densities and the drag effect is due to coupling of charge fluctuations in each ring . in a system with zero disorder , the relative magnitude of the drag current produced is typically a small fraction of the driving current . for two such systems separated by a distance @xmath46 the drag current is on the order of @xmath47 times the current in the driving system . @xcite based on the analysis of the last section , we expect that the drag current of two wigner crystals could in principle be as large as the driving current . the model of the one dimensional wigner crystal that we use here is that proposed in references ( @xcite ) and ( @xcite ) . in this model the wigner crystal is viewed as an elastic chain of spinless electrons . in the continuum limit , the wigner crystal of length @xmath48 , in the presence of an ahranov bohm flux , @xmath43 , is described by the lagrangian @xcite where @xmath12 is the crystal period , m is the electron mass , s is the velocity of sound in the crystal and @xmath50 is the flux quantum . the field variable @xmath51 describes the local displacement of the chain at a point @xmath52 at time @xmath32 . if the field variable is expanded in a fourier series @xmath53 we get , @xmath54 where @xmath55 . it is clear that the flux will only couple to the zero mode of the ring . so in dealing with the persistent current of a disorder free wigner crystal ring , the internal dynamics of the ring may be ignored and only the rotation of the ring as a whole is relevant . we wish to study the interaction between two such wigner crystal rings as shown in figure ( [ fig : two_rings ] ) . we take as a model for two disorder free interacting wigner crystal rings the following hamiltonian . @xmath56 where @xmath57 is the total mass of each ring , @xmath58 represents the strength of the inter - ring interaction , @xmath59 is the flux through ring 1 , @xmath60 is the flux through ring 2 and @xmath61 represents the displacement of ring @xmath62 respectively . the first two terms in equation ( [ eq : wc_ringl ] ) represent the individual dynamics of rings @xmath7 and @xmath9 respectively . the third term represents the interaction between the two rings . it is evident that the interaction between the chains will be periodic in the relative displacement between the two crystals . for simplicity we model this interaction as a cosine term , however the basic features of this analysis should not depend on the particular form of the periodic potential used . transforming to center of mass and relative angular coordinates @xmath63 and @xmath64 , we have in dimensionless form @xmath65\psi=\epsilon\psi\ ] ] where @xmath66 , @xmath67 , @xmath68 and @xmath69 . our task is now to solve equation ( [ eq : wc_ringh ] ) for the eigenenergy @xmath70 . the hamiltonian , equation ( [ eq : wc_ringh ] ) , separates into two parts @xmath71 with energies @xmath72 . the ground state wavefunction is a product @xmath73 . the equation for the center of mass coordinates is readily solved giving @xmath74 @xmath75 the equation for the relative coordinate wavefunction is @xmath76t(\phi)=0\ ] ] for @xmath77 this reduces to mathieu s equation . the energy eigenvalues correspond to the periodic solutions @xmath78 . we therefore substitute the fourier expansion @xmath79 into equation ( [ eq : relativeh ] ) and obtain the following recursion relation for the coefficients @xmath80 @xmath81 ^ 2)c_n + \frac{q}{2}c_{n-1}=0.\ ] ] the energy eigenvalues may be found numerically to any desired accuracy by truncating the set of equations for some @xmath82 and searching for the value of @xmath70 which gives a zero value for the appropriate determinant of coefficients . @xcite the energy eigenvalues determined in this way , combined with the results of equation ( [ eq : comh ] ) allow the currents in each ring to be determined from @xmath83 and @xmath84.@xcite . the results are shown in figure ( [ fig : wigner_jvsq ] ) . figure ( [ fig : wigner_jvsq ] ) shows that the drag current does approach the drive current in magnitude as the interaction @xmath58 is increased which is qualitatively similar to the classical result . for large values of the interaction strength , @xmath85 or @xmath86 the two wigner crystals have essentially the same current with each having half the current that one ring would have if isolated . in order to relate the interaction in equation ( [ eq : wc_ringl ] ) to the true coulomb interaction we approximate the amplitude @xmath58 as follows . we want to calculate the potential felt by an electron a distance @xmath4 from a finite periodic system of length @xmath48 and period @xmath12 . we model our system by taking a length @xmath48 of an infinite line of electrons centered at @xmath52 . this system is shown schematically in figure ( [ fig : amplitude_approx ] ) . the first term is a constant in @xmath52 therefore we ignore it . for small separation distances @xmath92 , higher @xmath93 terms may be ignored and we keep only the @xmath94 term . this yields finally @xmath95 \cos\left(\frac{2\pi}{a } x\right)\ ] ] we can now identify the amplitude @xmath58 in equation ( [ eq : wc_ringl])for a crystal with n electrons as @xmath96 @xmath97 as mentioned in the introduction for a one dimensional system , a long range interaction will lead to the formation of a wigner crystal@xcite . the long range nature of the coulomb force should be apparent for electron densities , @xmath98 where @xmath99 is the bohr radius for the particular ring material.@xcite for a typical mesoscopic ring radius of @xmath100 , crystal lattice constant @xmath101 and @xmath102 we have @xmath103 so the upper limit in the integral in equation ( [ eq : wc_amplitude ] ) may be taken to infinity which yields @xmath104 where @xmath105 is the zeroth order bessel function . this implies an exponential decrease in the amplitude of the interaction with increasing separation between the two wigner crystals . comparison with equation ( [ eq : lock_cond ] ) shows that the two crystals will be `` locked '' together at a distance @xmath106 . so two very low density perfect wigner crystals will be essentially locked together when separated by a distance of four times the lattice constant or less . we now consider the effects of impurities on the dynamics of the wigner crystal . in the presence of an impurity , the phase of the wigner crystal will be pinned . in this case the crystal ring can not rotate as a whole so the current as presented in the last subsection is not possible . it was shown , however , by rice et al . @xcite , that a new type of current is possible , namely the tunneling of solitons through the impurity barrier . it is the tunneling of solitons that we will investigate in this section . we are interested in knowing the ratio of the drag current in ring two , to the drive current in ring one . we therefore take as an assumption the presence of a current in ring one and look for the relative magnitude of the drag current in ring two in the presence of a single impurity . this system is shown schematically in figure ( [ fig : two_ringsphase ] ) . the first two terms in equation ( [ eq : lagrangewc ] ) are the lagrangians describing the dynamics of rings @xmath7 and @xmath9 respectively . the last term describes the interaction between the two rings . we follow the work of krive et al . @xcite and introduce the dimensionless parameter @xmath111 which characterizes the magnitude of the quantum fluctuations of the wigner crystal , and the parameter @xmath112 which characterizes the energy scale in a wigner crystal of length @xmath48 . we will be concerned with stiff crystals that are weakly fluctuating characterized by @xmath113 . the basic approach that we will use here is a semiclassical approximation used by larkin and lee @xcite in the context of charge density waves and elaborated by kleinert@xcite for a single particle tunneling through a barrier . in this approach the probability amplitude of a soliton tunneling through an impurity is given by @xmath114 where @xmath22 is the action associated with the tunneling trajectory obtained by minimizing the function @xmath115 , where @xmath116 is the lagrangian , equation ( [ eq : lagrangewc ] ) , with @xmath32 replaced by the imaginary time @xmath117 . this approach amounts to approximating the path integral of the propagator by its value at the classical trajectory.@xcite we now want to answer the following question : given a current in ring @xmath7 that causes a rotation of the phase by @xmath118 , what is the probability that the phase of ring @xmath9 will also rotate by @xmath118 ? it is apparent that there are three possiblities : here we have limited our consideration to single tunneling trajectories . in other words , each shift of the phase of each ring occurs due to a single tunneling event and ignore all multistep tunneling trajectories . thus we only consider strong coupling between rings , @xmath128 , and strong pinning @xmath129 . we now calculate the tunneling amplitudes for each case above . in case ( @xmath119 ) , the lagrangian in equation ( [ eq : lagrangewc ] ) reduces to @xmath130\ ] ] this is simply the problem of a single wigner crystal ring in the presence of a periodic pinning potential . the classical wave equation resulting from this lagrangian in real time , is @xmath131 this is the sine - gordon equation and it admits the exact soliton solution @xmath132\ ] ] where @xmath133 , @xmath134 and @xmath135 the @xmath136 sign corresponds to a soliton rotating in the clockwise direction , the @xmath137 sign corresponds to an antisoliton rotating in the counterclockwise direction and @xmath138 is the soliton / antisoliton velocity . the boundary conditions on such a solution are @xmath139 . a solution to this equation that matches this boundary condition is the soliton / antisoliton solution and has the form @xmath140.\ ] ] this solution has the following interpretation : at @xmath141 the phase of the crystal is @xmath142 , as @xmath32 increases , the phase at @xmath143 locally increases to @xmath118 creating a region around @xmath143 with phase @xmath118 . as @xmath32 increases this region with @xmath118 phase propagates symmetrically from @xmath143 at speed @xmath138 towards @xmath144 and @xmath144 until the entire ring is at @xmath145 . the lagrangian , equation ( [ eq : lagrangecase1 ] ) has been considered in the context of charge density wave tunneling by bogachek et al . @xcite , and there it is noted that the homogeneous ( space independent ) soliton solution is advantageous in the action against any spatially dependent solution for a finite system . the homogenous solution corresponds to a uniform rotation of the wigner crystal as a whole and has the form @xcite @xmath146 this solution describes the uniform increase of the phase of the entire ring from @xmath142 to @xmath118 in a time @xmath147 . the action associated with this solution is by equation ( [ eq : lagrangecase1 ] ) @xmath148^{1/2}\ ] ] in case ( @xmath122 ) , the relative difference between the two phases does not increase , while the phase of the center of mass of the two ring system increases by @xmath118 . we assume that during the tunneling process the two rings remain in phase to minimize the potential energy between them . if we rewrite equation ( [ eq : lagrangewc ] ) using @xmath149 and @xmath150 , we get the relevant lagrangian for case ( @xmath122 ) to be the case of soliton tunneling through an impurity in a strongly pinned charge density wave was first considered by larkin and lee @xcite . in their analysis , the tunneling process is broken into two stages . the first stage consists of a rapid tunneling of a length @xmath153 of the crystal through the impurity in a short time @xmath154 and leaves the crystal in a distorted state of high potential energy . the second stage consists of a slow relaxation of the crystal from this distorted state back into a state that minimizes its potential energy . it is evident that this solution interpolates between the initial state , @xmath158 for all @xmath52 at @xmath141 , and the final state @xmath159 for @xmath143 at @xmath160 . the second stage of the process takes place away from the impurity @xmath161 and obeys the equation of motion @xmath162 the solution to this stage of the tunneling should obey the periodic boundary conditions of the ring and match the solution for the first stage at @xmath163 . this problem has been considered in the work of krive et al @xcite where the proposed trajectory for the relaxation stage has the form @xmath164 the action associated with the tunneling stage of the process is @xmath165 where @xmath166 are constants . the action associated with the relaxation stage is @xmath167 the length @xmath153 is obtained by minimizing the total action @xmath168 with the result that @xmath169 and @xmath170 where @xmath171 is a constant and @xmath172 is a constant assumed to be of the order unity @xcite . equation ( [ eq : actioncase2a ] ) shows the interesting result that the action is only logarithmically dependent on the impurity for strong pinning . in case ( @xmath125 ) above , the center of mass of the two ring system does not move while the phase of the relative coordinates increases by @xmath173 . the relevant lagrangian for this system is given by @xmath174 - w\left[1-\cos\left(\frac{\phi}{2}\right)\right].\ ] ] this is identical to the lagrangian in equation ( [ eq : lagrangecase2 ] ) with the addition of the term proportional to @xmath175 . since the probability for this process goes as @xmath176 it is expected that the probability for this case will be reduced relative to that in case ( @xmath122 ) by @xmath177 and can therefore be neglected . it is clear that if @xmath179 the two crystal will essentially be locked together . for purposes of illustration we consider a stiff crystal with @xmath180 and consider values of @xmath181 for stiff pinning @xmath152 . figure ( [ fig : impurity_lock ] ) shows that the impurity effects are not that drastic in determining the value of @xmath175 where the locking of the two crystals occur . this is due to the fact that the impurity potential enters logarithmically into the tunneling action . comparison with figure ( [ fig : wigner_jvsq]b ) shows that the distance at which the two wigner crystals become `` locked '' together is essentially the same as in the impurity free case . the main effect of the impurity is then in the magnitude of the persistent current . we can estimate the magnitude of these currents by noting that if the two currents are locked together , then the magnitude of the persistent current for a given flux @xmath43 should be the same as a single wigner crystal ring with twice the mass . this current has been calculated at zero temperature to be @xcite @xmath182 we see that the current in equation ( [ eq : currentweakpot ] ) is much smaller than the disorder free persistent current due to the @xmath183 factor . in summary we have analyzed the coulomb drag between two one - dimensional wigner crystal rings . for sufficiently large interaction between rings the drag current is essentially equal in magnitude to the drive current . for an impurity free ring this `` locking '' of currents occurs at a separation distance of @xmath184 with @xmath12 the crystal period . we analyzed the effect of a single impurity in the drag ring and found that the two currents are `` locked '' at essentially the same distance and the major effect of the impurity is to significantly decrease the magnitude of the persistent current . j. r. clem _ phys . b _ * 12 * , 1742 ( 1975 ) ; j. w. ekin , and j. r. clem _ phys . rev . b _ * 12 * , 1753 ( 1975 ) k. k. uprety and d. dominguez _ phys . rev . b _ * 51 * , 5955 ( 1995 ) ; h.j . lett _ * 71 * , 1864 ( 1993 ) ; g. fano , f. ortolani , a. parola ans l. ziosi _ phys . rev . b _ * 60 * , 15654 ( 1999 ) ; a.l . ivanov and h. haug lett _ * 71 * , 3182 ( 1993 ) ;
we consider the coulomb drag between two metal rings in which the long range coulomb interaction leads to the formation of a wigner crystal . the first ring is threaded by an ahranov bohm flux creating a persistent current @xmath0 . the second ring is brought in close proximity to the second and due to the coulomb interaction between the two rings a drag current @xmath1 is produced in the second . we investigate this system at zero temperature for perfect rings as well as the effects of impurities . we show that the wigner crystal state can in principle lead to a higher ratio of drag current to drive current @xmath2 than in weakly interacting electron systems .
the realization of bose - einstein condensation ( bec ) of trapped atomic gases @xcite has generated in the last years a huge amount of experimental and theoretical research @xcite . bec was achieved by confining atoms in magnetic traps and lowering the temperature of the system via the evaporative cooling technique . in this technique , the hottest atoms are selectively removed from the system and the remaining ones rethermalize via two - body collisions . a common feature of these experiments is that the trapped gases are dilute ; mean - field theory then provides a useful framework to study the role of the interaction between particles . recently , a new feature has been added to the experiments : the magnetically trapped condensate is transferred into an optical lattice generated by interfering laser beams @xcite . with this new experimental setup it is possible to access the strongly correlated regime . the superfluid - mott - insulator transition was studied @xcite and other interesting physical phenomena such as the collapse and revival of the condensate have been found @xcite . the presence of the optical lattice and the fact that the particles interact only via contact interaction lead in a natural way to the hubbard model as a paradigm for these systems . a theoretical work proposing such experiments @xcite and recent quantum monte carlo ( qmc ) simulations @xcite have investigated these systems beyond the mean - field approximation . it has been found that the incompressible mott - insulating phase always coexists with compressible phases so that a local order parameter @xcite has to be defined to characterize the system . also , the phase diagram of trapped bosons was found to be more complex than the one in the homogeneous system @xcite . the experimental realization of ultracold fermionic gases is more difficult than the bosonic case , and has been achieved only recently @xcite . unlike bosons , single species fermions can not be directly evaporatively cooled to very low temperatures because the s - wave scattering that could allow the gas to rethermalize during the evaporation is prohibited for identical fermions . this problem has been overcome by simultaneously trapping and evaporatively cooling two - component fermi gases @xcite , and introducing mixed gases of bosons and fermions in which bosons enable fermions to rethermalize through their elastic interactions @xcite . more recently , rapid forced evaporation employing a feshbach resonance of two different spin states of the same fermionic atom has been used @xcite . now that it is possible to go well below the degeneracy temperature in the experiments and superfluidity appears within reach @xcite , it is expected that the metal - mott - insulator transition ( mmit ) could also be realized for fermions on an optical lattice . motivated by this expectation we study the ground state of the one - dimensional ( 1d ) fermionic hubbard model with a harmonic trap and with repulsive contact interaction , using qmc simulations . like in the bosonic case @xcite , mott - insulating domains appear over a continuous range of fillings and always coexist with compressible phases , such that global quantities are not appropriate to characterize the system . instead , we define a local - order parameter ( a local compressibility ) in order to characterize the local phases present in the system . we also analyze the generic features that are valid for any kind of confining potential , and not only for the harmonic one . it is well known that the hubbard model in the periodic case displays a mmit phase transition at half filling and at a finite value of the on - site repulsive interaction @xmath0 . ( in the case of perfect nesting the transition occurs at @xmath1 . ) the two routes to this mmit are the filling - controlled mmit and the bandwidth - controlled mmit @xcite . in the confined case , it is possible to drive the transitions by changing the total filling of the trap , the on - site repulsive interaction , or varying the curvature of the confining potential . in the following section , we study a number of local quantities as a function of the filling and the strength of the interaction . although they signal the appearance of a mott - insulating phase , such quantities provide no rigorous criterion to characterize the mott - insulating region . we therefore define a local compressibility that acts as a genuine local - order parameter . in sec . iii , we study the momentum distribution function for the confined system . in sec . iv , we discuss a mean - field approach for the 1d trapped system and compare the results with the ones obtained using qmc . in sect . v , we analyze the phase diagram and determine its generic form , which allows to compare systems with different sizes , number of particles and curvatures of the harmonic confining potential . in this section we also analyze the extension , for arbitrary confining potentials , of the results obtained for the harmonic case . finally , the conclusions are given in sec . some of the results presented here were summarized in a previous publication @xcite . due to the fact that the confining potential leads to an inhomogeneous density profile , we study in the present section how local quantities behave in the trapped system when the parameters at hand are changed . the hamiltonian of the fermionic hubbard model with a confining parabolic potential has the form @xmath2 where @xmath3 , @xmath4 are creation and annihilation operators , respectively , for a fermion with spin @xmath5 at site @xmath6 , and @xmath7 , such that @xmath8 is the hopping amplitude , @xmath0 is the on - site interaction that in the present work will be considered repulsive ( @xmath9 ) , @xmath10 is the curvature of the confining harmonic potential , and @xmath11 measures the position of the site @xmath6 ( @xmath12 with @xmath13 the lattice constant ) . the number of lattice sites is @xmath14 and is selected so that all the fermions are confined in the trap . we denote the total number of fermions in the trap as @xmath15 and consider equal number of fermions with spins up and down ( @xmath16 ) . in our simulations , we used the zero - temperature projector method @xcite adapted from the qmc determinantal algorithm by blankenbecler , scalapino , and sugar @xcite . the discrete hubbard - stratonovich transformation by hirsch was used @xcite . for details of the algorithm we refer to a number of reviews @xcite . the results for the evolution of the local density ( @xmath17 ) , as a function of the total number of the confined particles , are shown in fig . [ perfil3d ] . for the lowest filling , so that @xmath18 at every site , the density shows a profile with the shape of an inverted parabola , similar to that obtained in the non - interacting case @xcite , and hence , such a situation should correspond to a metallic phase . increasing the number of fermions up to @xmath19 , a plateau with @xmath20 appears in the middle of the trap , surrounded by a region with @xmath21 ( metallic ) . since in the homogeneous case , a mott insulator appears at @xmath20 , it is natural to identify the plateau with such a phase . the mott - insulating domain in the center of the trap increases its size when more particles are added , but at a certain filling ( @xmath22 here ) this becomes energetically unfavorable and a new metallic phase with @xmath23 starts to develop in the center of the system . upon adding more fermions , this new metallic phase widens spatially and the mott - insulating domains surrounding it are pushed to the borders . depending on the on - site repulsion strength , they can disappear and a complete metallic phase can appear in the system . finally , a `` band insulator '' ( i.e. , @xmath24 ) forms in the middle of the trap for the highest fillings ( after @xmath25 here ) . due to the full occupancy of the sites , it will widen spatially and push the other phases present in the system to the edges of the trap when more fermions are added . , @xmath26 and @xmath27 . the positions are measured in units of the lattice constant @xmath13.,scaledwidth=46.0%,scaledwidth=33.0% ] although the existence of flat regions in the density profile is an indication that there is an insulator there , a more quantitative characterization is needed . as shown in ref . @xcite , the variance of the local density ( @xmath28 ) may be used on a first approach ( from here on , we refer to the variance as the variance of the local density ) . in fig . [ perfiles150 ] we show four characteristic profiles present in fig . [ perfil3d ] and their respective variances when the number of fermions in the system are @xmath2950 ( a ) , 68 ( b ) , 94 ( c ) , and 150 ( d ) . for the case in which only a metallic phase is present [ fig.[perfiles150](a ) ] , it is possible to see that the variance decreases when the density approaches @xmath20 and has a minimum for densities close to that value . for the mott - insulating domain [ @xmath20 in fig . [ perfiles150](b ) ] , the variance has a constant value that is smaller than that of the metal surrounding it . as soon as the mott - insulating phase is destroyed in the middle of the trap and a new metallic region with @xmath30 develops there [ fig . [ perfiles150](c ) ] , the variance increases in this region . the variance in the metallic region with @xmath30 will start to decrease again when the density approaches @xmath24 and will have values even smaller than those in the mott - insulating phase . finally , when the insulator with @xmath24 is formed in the center of the trap the variance vanishes there [ fig . [ perfiles150](d ) ] . ) and their variances ( ) . the fillings are @xmath2950 ( a ) , 68 ( b ) , 94 ( c ) , and 150 ( d).,scaledwidth=49.0%,scaledwidth=41.0% ] alternatively , mott - insulating regions can be obtained by increasing the ratio between the on - site repulsive interaction and the hopping parameter as was done in experiments for bosons confined in optical lattices @xcite . [ perfdeltcomp100](a ) shows the evolution of the density profiles in a trapped system with @xmath31 , @xmath22 , and @xmath32 when this ratio is increased from @xmath33 to @xmath34 [ for more details of these density profiles see fig . [ perfilk100](e)-[perfilk100](h ) ] . it can be seen that for small values of @xmath35 ( @xmath33 ) there is only a metallic phase present in the trap . as the value of @xmath35 ( @xmath36 ) is increased , a mott - insulating phase tries to develop at @xmath20 while a metallic phase with @xmath30 is present in the center of the system . as the on - site repulsion is increased even further ( @xmath37 ) , a mott - insulating domain appears in the middle of the trap suppressing the metallic phase that was present there . in fig . [ perfdeltcomp100](b ) we show the variance of the density for the profiles in fig . [ perfdeltcomp100](a ) ( from top to bottom , the values presented are for @xmath38 ) . as expected , the variance decreases in both the metallic and mott - insulating phases when the on - site repulsion is increased . when the mott - insulating plateau is formed in the density profile , a plateau with constant variance appears in the variance profile with a value that will vanish only in the limit @xmath39 . as shown in fig . [ perfdeltcomp100](b ) , whenever a mott - insulating domain is formed in the trap , the value of the variance in it is exactly the same as the one for the mott - insulating phase in the homogeneous system for the same value of @xmath35 ( horizontal dashed lines ) . this would support the validity of the commonly used local density ( thomas - fermi ) approximation @xcite . however , the insets in fig . [ perfdeltcomp100](b ) , show that this is not necessarily the case , since for @xmath36 , the value of the variance in the mott - insulating phase of the homogeneous system is still not reached in the trap , although the density reaches the value @xmath40 . therefore , in contrast to the homogeneous case , a mott - insulating region is not determined by the filling only . in the cases of @xmath41 [ inset in fig . [ perfdeltcomp100](b ) for a closer look ] and @xmath34 , the value of the variance in the homogeneous system is reached and then we can say that mott - insulating phases are formed there . and @xmath22 , the on - site repulsions are @xmath422 ( ) , 4 ( ) , 6 ( ) , and 8 ( ) . ( a ) local density , ( b ) variance of the local density , ( c ) local compressibility @xmath43 as defined in eq . ( [ localc ] ) . the dashed lines in ( b ) are the values of the variance in the @xmath20 homogeneous system for @xmath44 ( from top to bottom).,scaledwidth=43.0%,scaledwidth=73.0% ] although the variance gives a first indication for the formation of a local mott insulator , an ambiguity is still present , since there are metallic regions with densities very close to @xmath45 and @xmath24 , where the variance can have even smaller values than in the mott - insulating phases . therefore , an unambiguous quantity is still needed to characterize the mott - insulating regions . we propose a local compressibility as a local - order parameter to characterize the mott - insulating regions , that is defined as @xmath46 where @xmath47 is the density - density correlation function and @xmath48 , with @xmath49 the correlation length of @xmath50 in the unconfined system at half - filling for the given value of @xmath0 . as a consequence of the charge gap opened in the mott - insulating phase at half filling in the homogeneous system , the density - density correlations decay exponentially [ @xmath51 enabling @xmath49 to be determined . the factor @xmath52 is chosen within a range where @xmath43 becomes qualitatively insensitive to its precise value . since there is some degree of arbitrariness in the selection of @xmath52 , we will examine its role in the calculation of @xmath43 here . for the homogeneous and trapped ( @xmath32 , @xmath22 ) systems in the mott - insulating and metallic phases , @xmath53 . ( ) @xmath20 homogeneous , ( ) @xmath54 trapped , ( ) @xmath55 homogeneous , and ( ) @xmath56 trapped . , scaledwidth=42.0%,scaledwidth=24.0% ] we show in fig . [ compvsl_qmc ] ( in a semi - log plot ) how the local compressibility behaves as a function of @xmath57 for a homogeneous and a trapped systems in the mott - insulating ( @xmath20 ) and metallic ( @xmath58 ) phases . in the mott - insulating phase , it is possible to see that in the homogeneous and confined systems the local compressibility decays exponentially to zero in exactly the same way due to the charge gap present there . in the metallic case there is no charge gap so that density - density correlations decay as a power law and the local compressibility , shown in fig . [ compvsl_qmc ] , decays very slowly . in fig . [ compvsl_qmc ] , a departure in the behavior of the local compressibility for the trapped system from the homogeneous case for large values of @xmath57 can be seen . this occurs because correlations with points very close to the band insulating ( @xmath45 ) and the mott - insulating ( @xmath20 ) regions surrounding the metallic phase [ see fig . [ perfdeltcomp100](a ) ] are included in @xmath59 . however , this effect does not affect the fact that the value of the local compressibility is clearly different from zero as long as the size of the metallic phase is larger than @xmath60 . we obtain that considering @xmath61 ( no matter what the exact value of @xmath52 is ) , the local compressibility is zero with exponential accuracy ( i.e. @xmath62 or less ) for the mott - insulating phase and has a finite value in the metallic phase . for the case in fig . [ compvsl_qmc ] where @xmath53 , the correlation length is @xmath63 so that @xmath64 . physically , the local compressibility defined here gives a measure of the change in the local density due to a constant shift of the potential over a finite range but over distances larger than the correlation length in the unconfined system . in fig . [ perfdeltcomp100](c ) , we show the profiles of the local compressibility for the same parameters as figs.[perfdeltcomp100](a ) and ( b ) ( we did not include the profile of the local compressibility for @xmath65 because for that value of @xmath0 we obtain that @xmath57 is bigger than the system size ) . in fig . [ perfdeltcomp100](c ) , it can be seen that the local compressibility only vanishes in the mott - insulating domains . for @xmath26 , it can be seen that in the region with @xmath66 the local compressibility , although small , does not vanish . this is compatible with the fact that the variance is not equal to the value in the homogeneous system there , so that although there is a shoulder in the density profile , this region is not a mott insulator . therefore , the local compressibility defined here serves as a genuine local order parameter to characterize the insulating regions that always coexist with metallic phases . at this point we would like to remark that , as shown in ref . @xcite , the local compressibility shows critical behavior on entering the mott - insulating region . this is the reason to speak about phases , since on passing from one region to the other , critical behavior sets in . finally , we discuss in this section the spin - spin correlation function , since in periodic chains , quasi - long - range antiferromagnetic correlations appear in the mott - insulating phase . in fig . [ spinesz ] we show the local @xmath67 correlation function for some points of the profiles presented in fig . [ perfiles150 ] . we measured the spin - spin correlations in the trap at the points in the figures where it can be clearly seen that @xmath68 has the maximum value . in fig . [ spinesz](a ) it can be seen that in the metallic phase with @xmath21 , the spin - spin correlations decay rapidly and do not show any clear modulation , which is due to the fact that the density is changing in this region . in the local mott - insulating phase [ center of the trap in fig . [ spinesz](b ) ] , short - range antiferromagnetic correlations are present , and they disappear completely only on entering in the metallic regions . for the shoulders with @xmath66 [ fig . [ spinesz](c ) ] , the antiferromagnetic correlations are still present but due to the small size of these regions they decay very rapidly . finally for the metallic regions with @xmath30 , the spin correlations behave like in the metallic phases with @xmath21 as shown in fig . [ spinesz](d ) . ( @xmath2950 ) ( a ) , @xmath69 ( @xmath2968 ) ( b ) , @xmath70 ( @xmath2994 ) ( c ) , and @xmath71 ( @xmath29150 ) ( d ) , for a trap with @xmath72 , @xmath26 , and @xmath27 ( the density and variance profiles for these parameters were presented in fig . [ perfiles150]).,scaledwidth=48.0%,scaledwidth=44.0% ] in most experiments with quantum gases carried out so far , the momentum distribution function , which is determined in time - of - flight measurements , played a central role . a prominent example is given by the study of the superfluid - mott - insulator transition @xcite in the bosonic case . also a qmc study relating this quantity to the density profiles and proposing how to determine the point at which the superfluid - mott - insulator transition occurs was presented in ref . as shown below , we find that this quantity is not appropriate to characterize the phases of the system in the fermionic case , and does not show any clear signature of the mmit . in fig . [ perfilk150 ] we show the normalized momentum distribution function ( @xmath73 ) for the same density profiles presented in fig . [ perfiles150 ] . for the trapped systems , we always normalize the momentum distribution to be unity at @xmath74 . in the case presented in fig . [ perfilk150 ] , transitions between phases occur due to changes in the total filling of the trap . we first notice that @xmath73 for the pure metallic phase in the harmonic trap [ fig . [ perfilk150](a ) ] does not display any sharp feature corresponding to a fermi surface , in clear contrast to the homogeneous case . the lack of a sharp feature for the fermi surface is independent of the presence of the interaction and is also independent of the size of the system . in the non - interacting case , this can be easily understood : the spatial density and the momentum distribution will have the same functional form because the hamiltonian is quadratic in both coordinate and momentum . when the interaction is present , it could be expected that the formation of local mott - insulating domains generates a qualitatively and quantitatively different behavior of the momentum distribution , like in the homogeneous case where in the mott - insulating phase the fermi surface disappears and @xmath73 is smoother . in fig . [ perfilk150](b ) , it can be seen that there is no qualitative change of the momentum distribution when the mott - insulating phase is present in the middle of the trap . quantitatively @xmath73 in this case is very similar to the pure metallic cases fig . [ perfilk150](a ) ( for densities smaller than one ) and fig . [ perfilk150](c ) ( where in the middle of the trap the density is higher than one ) . only when the insulator with @xmath24 appears in the middle of the system we find a quantitative change of @xmath73 for @xmath26 , as shown in fig . [ perfilk150](d ) . . the fillings are @xmath2950 ( a ) , 68 ( b ) , 94 ( c ) , and 150 ( d).,scaledwidth=48.0%,scaledwidth=44.0% ] we also studied the momentum distribution function when the mmit is driven by the change of the on - site repulsion . we did not observe any clear signature of the formation of the mott - insulating phase in @xmath73 . in fig . [ perfilk100 ] we show the normalized momentum distribution function [ fig . [ perfilk100](a)-[perfilk100](d ) ] for density profiles [ fig . [ perfilk100](e)-[perfilk100](h ) ] in which the on - site repulsion was increased from @xmath33 to 8 . it can be seen that the same behavior present in fig . [ perfilk150 ] and the quantitative changes in @xmath73 appear only when the on - site repulsion goes to the strong - coupling regime , but this is long after the mott - insulating phase has appeared in the system . ( a),(e ) , 4 ( b),(f ) , 6 ( c),(g ) , 8 ( d),(h ) and @xmath31 , @xmath22 , @xmath32.,scaledwidth=49.0%,scaledwidth=74.0% ] at this point one might think that in order to study the mmit using the momentum distribution function , it is necessary to avoid the inhomogeneous trapping potential and use instead a kind of magnetic box with infinitely high potential on the boundaries . however , in that case one of the most important achievements of the inhomogeneous system is lost , i.e. , the possibility of creating mott - insulating phases for a continuous range of fillings . in the perfect magnetic box , the mott - insulating phase would only be possible at half filling , which would be extremely difficult ( if possible at all ) to adjust experimentally . the other possibility is to create traps which are almost homogenous in the middle and which have an appreciable trapping potential only close to the boundaries . this can be studied theoretically by considering traps with higher powers of the trapping potentials . as shown below , already non - interacting systems make clear that a sharp fermi edge is missing in confined systems . in fig . [ perfilk1000](a ) we show the density profile of a system with 1000 sites , @xmath75 , and a trapping potential of the form @xmath76 with @xmath77 . it can be seen that the density is almost flat all over the trap with a density of the order of one particle per site . only a small part of the system at the borders has the variation of the density required for the particles to be trapped . in fig . [ perfilk1000](b ) ( continuous line ) , we show the corresponding normalized momentum distribution . it can be seen that a kind of fermi surface develops in the system but for smaller values of @xmath78 , @xmath73 is always smooth and its value starts decreasing at @xmath74 . in order to see how @xmath73 changes when an incompressible region appears in the system , we introduced an additional alternating potential , so that in this case the new hamiltonian has the form @xmath79 where @xmath80 is the strength of the alternating potential . for the parameters presented in fig . [ perfilk1000](a ) , we obtain that a small value of @xmath80 ( @xmath81 ) generates a band insulator in the trap , which extends over the region with @xmath66 ( when @xmath82 ) . however , the formation of this band insulator is barely reflected in @xmath73 , as can be seen in fig . [ perfilk1000](b ) ( dashed line ) . only when the value of @xmath80 is increased and the system departs from the phase transition [ @xmath83 , dotted line in fig . [ perfilk1000](c ) ] , does a quantitatively appreciable change in @xmath73 appear . noninteracting trapped fermions in a lattice with 1000 sites and a confining potential @xmath77 . density profile ( a ) and the normalized momentum distribution function ( b ) : the continuous line corresponds to ( a ) , the dashed line is the result when an alternating potential @xmath81 is superposed on the system , and the dotted line corresponds to @xmath83.,scaledwidth=40.0%,scaledwidth=47.0% ] in general , it is expected that on increasing the system size , the situation in the homogeneous system is approached . however , in the present case it is not merely a question of boundary conditions , but the whole system is inhomogeneous . therefore , a proper scaling has to be defined in order to relate systems of different sizes . in the case of particles trapped in optical lattices when there is a confining potential with a power @xmath84 and strength ( @xmath85 ) , a characteristic length of the system ( @xmath86 ) is given by @xmath87 , so that a characteristic density ( @xmath88 ) can be defined as @xmath89 . we find that this characteristic density is the one meaningful in the thermodynamic limit . in fig . [ perfilfree ] we show three systems in which the total number of particles and the curvature of the confining potential @xmath90 were changed , keeping @xmath88 constant . . ( a ) density profiles , ( b ) normalized momentum distribution function . dashed line corresponds to @xmath91 , @xmath92 ( @xmath93 ) , continuous line corresponds to @xmath77 , @xmath75 ( @xmath94 ) and dotted line corresponds to @xmath95 , @xmath96 ( @xmath97 ) . in the density profiles the positions are given in units of the characteristic length @xmath86.,scaledwidth=40.0%,scaledwidth=49.0% ] we measured the positions in the trap in units of the characteristic length @xmath86 . these systems have occupied regions ( @xmath98 ) with very different sizes , of the order of 500 lattice sites for the dashed line , of the order of 1000 sites for the continuous line , and of the order of 2000 lattice sites for the dotted line . [ perfilfree](a ) shows that the density profiles scale perfectly when the curvature of the confining potential and the number of fermions are changed in the system , so that in order to show the changes in the density profile , we introduced an inset that expands a region around a density 0.85 . in the case of the momentum distribution function [ fig . [ perfilfree](b ) ] , it is possible to see that this quantity also scales very well so that almost no changes occur in the momentum distribution function when the occupied system size is increased , implying that in the thermodynamic limit the behavior of @xmath73 is different from the one in the homogeneous system . an expanded view of the region with @xmath73 around 0.87 is introduced to better see the scale of the differences in this region . in sec . v we show that the characteristic density defined here is also the meaningful quantity to define the phase diagram of the system . the mean - field ( mf ) approach has been very useful in the study of dilute bosonic gases confined in harmonic potentials @xcite . this theory for the order parameter associated with the condensate of weakly interacting bosons for inhomogeneous systems takes the form of the gross - pitaevskii theory . when the bosonic condensate is loaded in an optical lattice , it is possible to go beyond the weakly interacting regime and reach the strongly correlated limit for which a superfluid - mott - insulator transition occurs . also in this limit a mf study was done @xcite , and the results were compared with exact diagonalization results for very small systems , reporting a qualitatively good agreement between both methods @xcite . in this section , we compare a mf approximation with qmc results for the system under consideration . it will be shown that the mf approach not only violates the mermin - wagner theorem , leading to long - range antiferromagnetic order in one dimension , as expected , but also introduces spurious structures in the density profiles . in order to obtain the mf hamiltonian , we rewrite the hubbard hamiltonian in eq . ( [ hubb ] ) in the following form ( up to a constant shift in the chemical potential ) : @xmath99 \notag \\ & & -\frac{\lambda } { 2 } [ \left\langle \widehat{\mu } _ { i}\right\rangle \widehat{\mu } _ { i}-\left\langle \widehat{\mu } _ { i}\right\rangle ^{2}+ \widehat{\delta } _ { \mu _ { i}}^{2 } ] , \label{mf1}\end{aligned}\ ] ] where the density and the magnetization in each site are denoted by @xmath100 and @xmath101 , respectively . the fluctuations of the density and magnetization are given by @xmath102 and @xmath103 , respectively , and @xmath104 is an arbitrary parameter that was introduced in order to allow for the most general variation in parameter space . the following relations for fermionic operators were used : @xmath105 neglecting the terms containing the square of the density and magnetization fluctuations in eq . ( [ mf1 ] ) , the following mf hamiltonian @xmath106 \hat{n}_{i\uparrow } \notag \\ & & + \frac{1}{2 } [ \left ( 1-\lambda \right ) \left\langle \hat{n } _ { i}\right\rangle + \lambda \left\langle \widehat{\mu } _ { i}\right\rangle ] \hat{n}_{i\downarrow } \notag \\ & & -\frac{\left ( 1-\lambda \right ) } { 2}\left\langle \hat{n}_{i}\right\rangle ^{2}+\frac{\lambda } { 2}\left\langle \widehat{\mu } _ { i}\right\rangle ^{2 } \label{mf2}\end{aligned}\ ] ] is obtained . due to the inhomogeneity of the trapped system , an unrestricted hartree - fock scheme is used to determine the local densities @xmath107 and local magnetic moments @xmath108 . given a mf ground state @xmath109 , the minimum energy @xmath110 ( not the mf one @xmath111 ) is reached when @xmath112 , so that all the results that follow were obtained for this value of @xmath104 . we first recall some of the discrepancies between mf approximations and the known valid facts for the 1d homogenous system ( @xmath113 ) as follows . \(i ) at half filling it is known that the hubbard model exhibits a mott - insulating phase with a charge gap while the spin sector remains gapless , so that the density - density correlations decay exponentially and the spin - spin correlations decay as a power law . within the mf approximation , there is a band insulator in the system at half filling so that both the density - density and spin - spin correlations decay exponentially . the mf value of the charge gap is an overestimation of the real one , as can be seen in fig . [ correlhomf ] where we present the mf results for the density - density correlations as function of the distance , together with the qmc results . at half filling for @xmath53 . mf ( ) and qmc ( ) solutions.,scaledwidth=40.0%,scaledwidth=21.0% ] the slopes of the curves are proportional to the charge gap and the mf slope is approximately twice the qmc one ( the results were obtained for a system with 102 sites and @xmath53 ) . the correlation length of @xmath114 is the inverse of these slopes . finally , at half filling the mf theory leads to an antiferromagnetic state , while in the hubbard model the magnetization is always zero although quasi - long - range antiferromagnetic correlations appear in the system . \(ii ) at any noncommensurate filling , the hubbard model in 1d describes a metal ( luttinger liquid ) , so that no gap appears in the charge and spin excitations , and there is a @xmath115 modulation in the spin - spin correlation function that leads to the well - known @xmath116 singularity . within the mf approximation , there is always a band insulator at any noncommensurate filling , with a gap that decreases when the system departs from half filling . the appearance of this gap for any density is due to the perfect nesting present in one dimension . the insulating nature of these solutions can be seen in the global compressibility of the system that is always zero , in the behavior of the density - density and the spin - spin correlations which decay exponentially , and in the momentum distribution function where there is no fermi surface , as shown in fig . [ perfilk102o066mf ] for @xmath117 , @xmath118 , and @xmath53 . , @xmath118 , and @xmath53.,scaledwidth=39.0%,scaledwidth=23.0% ] within this mf approach , there is a @xmath116 modulation of the magnetization only in the @xmath119 component [ the @xmath120 symmetry was broken ] , which leads to a divergence of the fourier transform of @xmath121 at @xmath122 . in fig . [ spinesk102o066 ] we compare the mf result ( a ) for @xmath123 with the qmc one ( b ) , for @xmath117 , @xmath118 , and @xmath53 , where it can be seen that the @xmath116 peak is one order of magnitude bigger in the mf case compared to the qmc case . this is due to the existence of the magnetization in the mf solution ( the values at the peaks will only diverge in the thermodynamic limit ) . in a homogeneous system with @xmath117 , @xmath118 , and @xmath53.,scaledwidth=49.0%,scaledwidth=24.0% ] we consider next the confined inhomogeneous case . in fig . [ perfdelt100mf](a ) we show the density profiles obtained for a trap with @xmath31 , @xmath22 , and @xmath32 ( like the one presented in fig . [ perfdeltcomp100 ] ) for different values of the on - site repulsion @xmath124 [ for more details see figs . [ perfilk100o070mf](e)-[perfilk100o070mf](h ) ] . it can be seen that the mf solutions have also density profiles in which `` metallic '' ( @xmath125 ) and insulating ( @xmath20 ) phases coexist . however , the mf insulating plateaus with @xmath20 [ @xmath36 in fig . [ perfdelt100mf](a ) ] appear for smaller values of @xmath0 than the ones required in the hubbard model for the formation of the mott - insulating plateaus [ @xmath41 in fig . [ perfdeltcomp100](a ) ] . in the `` metallic '' phases it is possible to see that the charge density shows rapid spatial variations that are very large for @xmath126 [ see also fig.[perfilk100o070mf](f ) ] . these density variations in the `` metallic '' phases are reflected in the variance profiles [ fig . [ perfdelt100mf](b ) ] . the values of the variance in the plateaus are the same as the ones obtained in the homogeneous mf case for the same value of @xmath0 when @xmath20 [ fig . [ perfdelt100mf ] ( b ) ] . this feature was discussed in sec . ii for the qmc solutions . , @xmath32 , and @xmath22 , the on - site repulsions are @xmath33 ( ) , 3 ( ) , 4 ( ) , and 6 ( ) . ( a ) local density , ( b ) variance of the local density . the dashed lines in ( b ) are the mf values of the variance in the @xmath20 homogeneous system for @xmath127 ( from top to bottom).,scaledwidth=48.0%,scaledwidth=57.0% ] since the mf solution leads to an insulating state for incommensurate fillings in the homogeneous case , we analyze in the following more carefully the regions with @xmath128 in the presence of the trap . we find that the local compressibility always decays exponentially as a function of @xmath57 over the entire system , although the exponents are different depending on the density of the point analyzed . some additional modulation appears in @xmath129 when @xmath128 , as shown in fig . [ compvsl_mf ] where the local compressibility is displayed as a function of @xmath57 for @xmath20 and @xmath58 in the trapped and homogeneous cases . therefore , although the mf approximation leads to a density profile similar to the one obtained in qmc as long as @xmath21 , it gives a qualitative wrong description of the character of those regions . for the homogeneous and trapped ( @xmath32 , @xmath22 ) systems for @xmath20 and @xmath130 , @xmath53 . ( ) @xmath20 homogeneous , ( ) @xmath54 trapped , ( ) @xmath130 homogeneous , and ( ) @xmath131 trapped . , scaledwidth=44.0%,scaledwidth=27.0% ] the mf results for the normalized momentum distribution function are presented in figs . [ perfilk100o070mf](a)-[perfilk100o070mf](d ) with their corresponding density profiles [ figs . [ perfilk100o070mf](e)-[perfilk100o070mf](h ) ] , for the same parameter values as in fig . [ perfdelt100mf ] . it can be seen that the normalized momentum profiles are similar to the ones obtained with qmc when the density profiles are similar , so that the mf results are very similar to the qmc ones for this averaged quantity although the physical situation is very different . ( within mf there is always an insulator in the system and within qmc there are metallic and insulating phases coexisting . ) in figs . [ perfilk100o070mf](e)-[perfilk100o070mf](h ) , we also show the @xmath132 component of the spin on each site of the trap [ @xmath133 . when the density is around @xmath20 , it can be seen that antiferromagnetic order appears , and that a different modulation exists for @xmath134 . we should mention that we have also found some mf solutions for the trapped system where there was formation of spin domain walls in the insulating regions with @xmath20 and there the variance was not constant ( a plateau in the density does not correspond to a plateau in the variance ) , so that care should be taken with the mean field solutions . ( ) profiles ( e)-(h ) for @xmath33 ( a),(e ) , 3 ( b),(f ) , 4 ( c),(g ) , 6 ( d),(h ) and @xmath31 , @xmath22 , @xmath32.,scaledwidth=77.0%,scaledwidth=93.0% ] in the present section , we study the phase diagram for fermions confined in harmonic traps with an underlying lattice . as it can be inferred from fig . [ perfil3d ] , the phases present in the system are very sensitive to the values of the parameters in the model , i.e. , to the curvature of the parabolic potential ( @xmath10 ) , to the number of fermions present in the trap ( @xmath15 ) and to the strength of the on - site repulsion ( @xmath0 ) . in order to be able to relate systems with different values of these parameters and different sizes , we use the characteristic density ( @xmath88 ) defined in sec . iii for the definition of the phase diagram . this quantity has been shown to be a meaningful quantity in the thermodynamic limit since density profiles ( as function of @xmath135 ) and normalized momentum distributions do not change when @xmath88 is kept constant and the filling in the system ( or the occupied system size ) is increased . for a harmonic potential , the characteristic length is given by @xmath136 and the characteristic density is then @xmath137 . in fig . [ dfs_scalingf ] we show the phase diagrams of two systems with different strengths of the confining potential ( @xmath138 and @xmath27 ) and different sizes ( @xmath31 and @xmath72 , respectively ) . this shows that the characteristic density is a meaningful quantity to characterize a generic phase diagram . it allows to compare different systems , and , hence , to relate the results of numerical simulations with larger experimental sizes ( although currently the linear dimension of the experiments is @xmath139 ) . ( ) and @xmath140 ( ) . the different phases are explained in the text.,scaledwidth=40.0%,scaledwidth=29.0% ] the phases present in the phase diagram in fig.[dfs_scalingf ] can be described as follows : ( a ) a pure metal without insulating regions , ( b ) a mott insulator in the middle of the trap always surrounded by a metal , ( c ) a metallic region with @xmath30 in the center of the mott - insulating phase , that is surrounded by a metal with @xmath21 , ( d ) an insulator with @xmath24 in the middle of the trap surrounded only by a metallic phase , and finally ( e ) a `` band insulator '' in the middle of the trap surrounded by metal with @xmath30 , and these two phases inside a mott insulator that as always is surrounded by a metal with @xmath21 . as it can be seen , this phase diagram is more complex than the one for the 1d homogeneous hubbard model . there are some features that we find interesting in this phase diagram : ( i ) for all the values of @xmath0 that we have simulated , we see that the lower boundary of phase b with the phase a has a constant value of the characteristic density . ( ii ) we have also seen that the upper boundary of phase b with phases a and c and the boundary of phases d and e with a and c , respectively , are linear within our errors . ( iii ) finally , there is another characteristic of phase b , and of the metallic phase a that is below phase b , that we find intriguing and could be related to point ( i ) . as can be seen in fig . [ perfdeltcomp100](a ) for @xmath41 and 8 , once the mott - insulating phase is formed in the middle of the trap , a further increase of the on - site repulsion leaves the density profile almost unchanged . ( the two mentioned profiles are one on top of the other in that figure . ) however , if we look at the variance of these profiles it decreases with increasing @xmath0 , which means that the double occupancy @xmath141/2 $ ] is decreasing in the system when @xmath0 is increased , without a redistribution of the density . less expected for us is that this also occurs for the metallic phase a that is below phase b. in fig.[perfdelta100o030](a ) we show four density profiles and in fig . [ perfdelta100o030](b ) their variances , for a trap with @xmath138 and @xmath142 when the on - site repulsion has values @xmath33 , 4 , 6 , and 8 . if we look at the phase diagram , the point with @xmath33 is not below phase b whereas the other three values are . in fig . [ perfdelta100o030](a ) we can see that as the value of @xmath35 is increased from 2 to 4 there is a visible change in the density profile , but after a further increase the profile remains almost the same for the other values of @xmath0 . the inset shows a magnification of the top of the profiles for a more detailed view . [ perfdelta100o030](b ) shows that although the density stays almost constant for @xmath143 , and 8 , the variance decreases . in the other phases ( c , d , and e ) , an increase of the on - site repulsion always changes the local densities pushing particles to the edges of the trap . and @xmath142 , the on - site repulsions are @xmath33 ( ) , 4 ( ) , 6 ( ) , and 8 ( ) . ( a ) local density , ( b ) variance of the local density.,scaledwidth=47.0%,scaledwidth=49.0% ] finally , in order to see , whether the features of the phase diagram discussed previously are particular of a harmonic confining potential , we have also performed an analogous study for a quartic confining potential . in fig . [ perfil3dv4 ] we show how the density profiles evolve in a trap with a quartic confining potential ( @xmath144 ) when the total filling is increased from @xmath145 to 140 . it can be seen that the shape of the metallic regions is flatter than in the harmonic case . however , all the features discussed previously for the parabolic confining potential are also present in the quartic case . local phases appear in the system , the mott - insulating plateaus with @xmath20 have values of the variance equal to the ones in the homogeneous case for the same value of @xmath0 , the local compressibility vanishes in the mott - insulating regions and the spin and momentum distribution exhibit a behavior similar to that of the parabolic case . , @xmath146 , and @xmath144.,scaledwidth=46.0%,scaledwidth=31.0% ] in addition , the form of the phase diagram for the quartic confining potential is similar to that of the harmonic trap . in fig . [ dfs_scalingv4f ] we show the phase diagram for a system with @xmath31 and @xmath144 , where the characteristic density is given by @xmath147 . we have also calculated some points of the phase diagram for another system with @xmath148 and @xmath31 in order to check that the scaling relation also works properly in this case . the phases are labeled in the same way as in fig . [ dfs_scalingf ] , up to the largest value of @xmath0 that we have studied we do not obtain the phase e which for the quartic confining case seems to be moved further towards the strong coupling regime . figure [ dfs_scalingv4f ] shows that the phase diagram in this case is very similar to that of the parabolic case , so that all the features discussed previously apply to the quartic confining potential . the form of the phase diagram in fig . [ dfs_scalingf ] seems to be of general applicability for any confining potential when the proper characteristic density is considered . ( ) and @xmath149 ( ) . the phases are labeled in the same way as in fig . [ dfs_scalingf].,scaledwidth=39.0%,scaledwidth=28.0% ] we have studied the ground - state properties of the 1d fermionic hubbard model in harmonic traps with an underlying lattice , using qmc simulations and a mf approach . due to the inhomogeneous density distribution , in general metallic and insulating regions coexist in the trap . therefore , local quantities have to be used to characterize the system . we considered first the variance of the local density , which , when a plateau develops in the density profile with @xmath20 , equals the value in the homogeneous system in the mott - insulating phase . however , it is not an unambiguous quantity , since it can be larger in the mott - insulating phases than in metallic regions with densities close to @xmath45 or @xmath24 . therefore , we have defined a local - order parameter ( local compressibility ) that vanishes in the insulating regions and is finite in the metallic ones . the local compressibility gives a measure of the local change in density due to a constant shift of the potential over finite distances larger than the correlation length of the density - density correlation function for the mott - insulating phase in the unconfined system . as expected , antiferromagnetic correlations are present in the mott - insulating phases and they remain in the shoulders of the density profile for @xmath66 . as opposed to homogeneous or periodic systems , where the appearance of a gap can in general be seen by the disappearance of the fermi edge in the momentum distribution function , it is easily seen that due to the presence of a confining potential , such a feature is much less evident . although increasing the power of the confining potential sharpens the features connected with the fermi edge , it remains qualitatively different from the homogeneous case . increasing the system size with a proper definition of the thermodynamic limit , does not change at all the behavior of the density and momentum distribution functions in the trapped case , such that , in fermionic systems it does not seem to be the appropriate quantity to look for evidence of gaps opening in the system . the mf study has shown that within this approach the trapped system is an insulator for all the densities in the trap , a qualitatively wrong picture in contradiction with the qmc results . in addition , the metallic regions showed large oscillations in the density that are not present in the real solutions . however , it is possible to obtain some qualitative information from mf , such as the coexistence of @xmath128 regions with @xmath20 plateaus , the shape of the momentum distribution function and the existence of antiferromagnetic correlations in the insulating regions with @xmath20 . finally , we have determined a generic form for the phase diagram that allows to compare systems with different values of all the parameters involved in the model . it can be used to predict the phases that will be present in future experimental results . the phase diagram also reveals interesting features such as reentrant behavior in some phases when some parameters are changed and phase boundaries with linear forms . a similar form of the phase diagram has been also found for a quartic confining potential , such that the results obtained here are generally applicable to cases beyond a perfect harmonic potential . we gratefully acknowledge financial support from the lfsp nanomaterialien . we are grateful to g. g. batrouni and r. t. scalettar for interesting discussions at the early stages of the project . we are grateful to f. f. assaad for valuable discussions about the mf approach we used here and also for providing us the core of the code we employed in our mf calculations . we thank r. noack and g. modugno for useful comments on the manuscript , and m. feldbacher and t. pfau for useful conversations . we thank hlr - stuttgart ( project dynmet ) for allocation of computer time , the calculations were carried out on the hitachi sr8000 . a. muramatsu , in _ quantum monte carlo methods in physics and chemistry _ , vol . 525 of nato advanced studies institute , series c : mathematical and physical sciences , edited by m.p . nightingale and c.j . umrigar , ( nato science series , kluwer academic , 1999 ) , pp . 343 - 373 . f. f. assaad , in _ quantum simulations of complex many - body systems : from theory to algorithms _ , edited by j. grotendorst , d. marx , and a. muramatsu ( john von neumann institute for computing ( nic ) , 2002 ) , vol . 10 , pp .
using quantum monte carlo ( qmc ) simulations we study the ground - state properties of the one - dimensional fermionic hubbard model in traps with an underlying lattice . since due to the confining potential the density is space dependent , mott - insulating domains always coexist with metallic regions , such that global quantities are not appropriate to describe the system . we define a local compressibility that characterizes the mott - insulating regions and analyze other local quantities . it is shown that the momentum distribution function , a quantity that is commonly considered in experiments , fails in giving a clear signal of the mott - insulator transition . furthermore , we analyze a mean - field approach to these systems and compare it with the numerically exact qmc results . finally , we determine a generic form for the phase diagram that allows us to predict the phases to be observed in the experiments .
there are two classes of models proposed for the formation of massive stars ( e.g. , krumholz 2005 ) . the first posits that massive stars are assembled in a manner analogous to their lower mass counterparts : from infalling material located in a rotating protostellar core channeled starward through a circumstellar accretion disk , with the flow of material from the inner regions of the accretion disk toward the stellar surface mediated by a magnetic field rooted in the star . we choose the shorthand magnetospherically - mediated accretion " ( henceforth mma ) to describe this scenario . bonnell , vine , & bate ( 2004 ) propose the following variant : that massive stars form preferentially in cluster - forming environments , starting their lives as lower mass cores whose mass then increases as a result of competitive accretion of surrounding molecular gas . in the case of mass buildup through competitive accretion , transfer of material from the protostellar core to the central forming star would presumably still follow the general precepts of mma . a second , very different class of model posits that massive stars start their lives as lower mass protostellar cores in clusters , and then grow via mergers of cores located in regions where the density of cores is high ( bally & zinnecker 2005 and references therein ) . we refer to this class of model as core merger " models or cm . observational tests aimed at determining which of these contrasting models , mma or cm , dominates in nature have proven impossible to date , primarily because high mass stars form on very short time scales in highly obscured regions at distances large enough so that , with the spatial resolutions achieved by current generation telescopes , crowding renders study of individual forming objects difficult . direct tests of mma for massive stars would involve measurements similar to those carried out to establish mma as the likely assembly mechanism for low mass stars : of stellar magnetic fields ( e.g. johns - krull , valenti , & koresko 1999 ) , and of line profiles diagnostic of the temperature - density - velocity fields expected were material channeled starward from the inner regions of the accretion disk by magnetic fields along funnel flows " ( e.g. , hartmann , hewett , & calvet , 1994 ; muzerolle , calvet , & hartmann 1998 ) . unfortunately , rotational and turbulent broadening among higher mass stars precludes direct observation of zeeman broadening for magnetically sensitive lines ( rotational broadening typically exceeds zeeman broadening by more than a factor of 10 ) . moreover , searches for line profile morphologies consistent with magnetic funnel flows have yielded no convincing evidence of such flows for stars with masses much in excess of 3 m@xmath0 ( e.g. , muzerolle et al . an alternative , but more indirect test of mma might be provided by observation of highly collimated outflows presumably launched in the inner regions of the accretion disk , perhaps near the point where the dynamical pressure of accretion is matched by the magnetic pressure of the stellar field ( e.g. , shu et al . such collimated outflows are common among low mass stars ( @xmath5 3 m@xmath0 ) which exhibit funnel flow profiles and for which direct measurements of magnetic field strengths are available , that is , objects where mma finds strong observational support . collimated outflows have also been observed in association with even more massive ( up to @xmath310 m@xmath0 ) stars surrounded by circumstellar accretion disks ( cesaroni et al . 2005 ; beuther & shepherd 2005 , and references therein ) , perhaps suggesting that mma is operative for stars as massive as 10 m@xmath0 . bally & zinnecker ( 2005 ) propose two possible tests for cm . the first involves surveys aimed at detecting direct evidence of collisions and mergers of low mass protostellar cores via observation of infrared flares " . the second is indirect , namely a change in the character of stellar outflows from the highly collimated outflows characteristic of lower mass objects ( @xmath6 10 m@xmath0 ) presumably undergoing mma , to much more poorly collimated outflows among higher mass objects . unfortunately , no surveys aimed at detecting flares in regions of high mass star formation have yet been carried out . moreover , there is thus far only very limited evidence regarding the degree of collimation , or lack thereof , for outflows from forming stars with masses larger than @xmath3 10 m@xmath0 . given the challenge of distinguishing observationally between mma and cm models for high mass star formation , alternative clues are potentially valuable . one such clue may be provided by observation of stellar angular momenta among young , high mass stars . for example , bally & zinnecker ( 2005 ) suggest that the product of the merger of two stars should be rotating rapidly : a result of converting the orbital angular momenta of merging protostars into the spin angular momentum of the resulting higher mass , merged object . in contrast , stars formed via mma should be rotating relatively slowly , at speeds well below breakup ( e.g. , shu et al . 1994 ) . in a recent paper discussing the rotation properties among young stars in orion , wolff , strom , & hillenbrand ( 2004 ; hereafter wsh ) used observations of @xmath1 to derive values of the projected specific angular momentum ( jsini / m ) for pre - main - sequence stars still on convective tracks ; these stars have recently completed their main accretion phase and should provide the best directly measurable estimate of the initial values of angular momentum . wsh found that the upper envelope of jsini / m varies slowly with mass ( j / m @xmath3 @xmath7 ) over the mass range 0.1 - 3 m@xmath0 and lies well below the angular momentum associated with stars having rotation speeds near breakup ( see section 3 below ) . these authors suggest that mma can in principle account for these results for stars in this mass range . in this paper we extend measurements of @xmath1 to o and b0 stars in extremely young clusters and associations and ask whether there is a discontinuity in rotation properties between lower mass stars and stars with masses @xmath8 3 m@xmath0 , as might be expected if mma dominates the formation of low mass stars , while cm plays a role in the formation of a significant fraction of high mass stars . the o stars for this study were selected from the associations studied by massey & thompson ( 1991 ) , hillenbrand et al . ( 1993 ) , and massey , johnson , & degioia - eastwood ( 1995 ) . because our proposed search for a discontinuity in rotation properties as a function of stellar mass requires our estimating rotation rates for stars as soon as possible after they have formed , we have limited our sample to stars very close to the zero - age main sequence ( zams ) , i.e. , to luminosity classes iii - v . the stars were observed with the hydra multi - object spectrograph at the wiyn telescope on kitt peak . observations ( the `` high resolution sample '' ) of nine o stars in ngc 6611 were obtained in may , 2005 using the hydra multiobject fiber spectrograph on wiyn as part of a more extensive study of stellar rotation in this cluster . in conjunction with an echelle grating and an order separating filter , these observations with wiyn - hydra yielded spectra with r @xmath3 20,000 spanning the wavelength range 4450 to 4590 @xmath9 . in addition to our ngc 6611 program stars , 12 stars with spectral types from o4 to b0 with known values of @xmath1 ( penny 1996 ; howarth et al . 1997 ) were observed on the same nights with the same instrumental configuration to serve as rotation standard stars . our approach was to measure the fwhm of he i 4471 @xmath9 and to derive a relationship between fwhm and published @xmath1 values ( penny 1996 ; howarth et al . 1997 ) for the twelve standard stars . we note that the @xmath1 values adopted from these two published studies result from application of a cross - correlation technique to the rich near - ultraviolet metallic spectra of late - o and early b- stars observed with iue . by using stars from these two previous studies as standards , we are placing our own values of @xmath1 on a known system . we note that different calibrations yield slightly different answers . for examples , the penny and howarth et al . values are larger by about 10% than the values determined from visual examination of observed line profiles by conti & ebbets ( 1977 ) from spectra in the optical region of the spectrum . in turn , these latter values are systematically smaller by about 30 km / sec than the values derived in the pioneering study of slettebak ( 1956 ) . forty - two additional stars were observed over the wavelength range 4070 - 4580 @xmath9 at a resolution of @xmath3 0.75 @xmath9 ( `` low resolution sample '' ) . these observations enable measurement of projected rotational velocities , @xmath1 @xmath10 50 km / sec , and upper limits to @xmath1 for stars rotating more slowly than 50 km / s . the spectra were extracted and calibrated in wavelength by kim venn . the lines used to measure @xmath1 were he i 4471 @xmath9 or he i 4387 @xmath9 or both , depending on the spectral type and the strength of the lines . in order to calibrate @xmath1 for our `` low resolution sample '' , we established the relationship between fwhm and @xmath1 for he i 4471 @xmath9 and he i 4387 @xmath9 by using the @xmath1 values determined as above from the 9 o stars in ngc 6611 that were observed at both resolutions . to strengthen the calibration , we also added 4 stars in ngc 6611 with spectral types b0.5 that were also observed at both high and low spectral resolutions ; the values of @xmath1 for these 13 stars range from 30 to 400 km / sec . ( the measurements of stars in ngc 6611 with types b0.5 and later will be reported along with data for b stars in several other associations in a subsequent paper . ) independent measures of the line widths for stars observed more than once during our run indicate that the values of @xmath1 are internally consistent to within about 20 km / sec . given the uncertainties in the calibration and the difficulty of measuring extremely broad lines , we report the overall uncertainty in a measured value of @xmath1 as @xmath310% or 20 km / sec , whichever is larger . this uncertainty is typical of @xmath1 studies for stars in this spectral type range . for example , the average difference in the @xmath1 values measured by both penny ( 1996 ) and howarth et al . ( 1997 ) for the 9 of our standard stars that are common to both of the published studies is 19 km / sec . as a further check on our values , we have compared our results with 11 stars for which data are also available on either webda ( http://www.univie.ac.at/webda/webda.html ) or from huang & gies ( 2006 ) . we find that the standard deviation of the difference between our results and the literature values is 13 km / sec . the stars observed , along with spectral types , etc . , are listed in table 1 . the first column lists the name of the cluster or association , the second gives the name of the star , if any ; the third lists the number on the webda system ; the fourth and fifth columns give the stellar position ; the sixth lists the spectral type , the seventh the log of the effective temperature , the eighth the bolometric luminosity , the ninth the stellar mass , the tenth the apparent rotational velocity , and the eleventh the log of the specific angular momentum . the derivation of the quantities in the last 5 columns is described in section 3.1 in order to make our first assessment of whether forming stars of all masses rotate similarly or whether there is a discontinuity at some characteristic mass , we need to establish a simple procedure for comparing rotation properties of young high and low mass stars . our first approach is to compare the observed rotational velocities of the stars in our sample with the critical velocity for stars of the same mass on the birthline . we will argue below that the observed values of @xmath1 for the very young stars in our sample have changed little since the time that the main accretion phase ended , and so the current values of @xmath1 are representative of their initial values . the goal of this comparison is to determine whether the ratio of the observed initial values of @xmath1 to the critical velocities on the birthline varies systematically with mass . this simple approach is motivated by the fact that explicit calculation of stellar angular momentum , which is a more fundamental quantity , requires a number of assumptions about stellar structure and the internal distribution of angular momentum as well as accurate estimates of stellar parameters . the apparent rotational velocity , however , is a directly measured quantity , while the breakup velocity on the birthline can be estimated from stellar models . therefore , we begin the analysis by comparing the observed values of @xmath1 with the critical velocities calculated from models of newly formed stars . models predict that stars with masses comparable to the o - stars in our sample are deposited directly on the zero age main sequence at the end of the assembly process . for a time - averaged mass accretion rate of @xmath11 = 10@xmath12 m@xmath0/year , which observations suggest is typical for intermediate mass stars ( palla & stahler 1992 ) , stars with m @xmath10 @xmath13 m@xmath0 should lie on the main sequence when accretion stops . if @xmath14 = 10@xmath15 m@xmath0/year , which is more typical of the accretion rates advocated by mckee & tan ( 2003 ) for 20 - 30 m@xmath0 stars , then stars with @xmath16 15 m@xmath0 should lie on the main sequence when accretion stops . therefore , we expect that all of the o stars in our sample were already on the main sequence when accretion ended . for condensed stasr , such as o stars on the main sequence , the breakup velocity @xmath17 is given by ( townsend , owocki , & howarth 2004 ) @xmath17 = @xmath18 , where m is the mass of the star and @xmath19 is the polar radius at the breakup velocity . the factor 2/3 takes into account the difference in the polar and equatorial radii for the case of critical rotation . for the o stars , most authors calculate the critical velocity by including a correction factor for radiation pressure : @xmath20 = @xmath21 where @xmath22 , the ratio of the luminosity to the eddington luminosity is given by mihalas ( 1978 ) : @xmath22 = 2.6 x @xmath23 l / m . maeder and meynet ( 2000 ) point out that this relationship is valid only if we assume that the brightness of a rotating star is uniform over its surface , which is inconsistent with von zeipel s theorem . they argue that radiation pressure can in fact be ignored for stars with eddington factors less than 0.639 , which is true for the stars treated here . for present purposes , we will also assume that the polar radius of a rapidly rotating star can be taken to be approximately equal to the radius of a non - rotating star . we have used the models of schaller et al . ( 1992 ) to obtain the parameters for zams o stars . the critical velocities calculated with no allowance for radiation pressure agree well with the calculations by meynet et al . ( 2006 ) . for an accretion rate of @xmath23 m@xmath0/yr , stars with masses less than 8 m@xmath0 have not yet reached the zams when the main accretion phase ends . to estimate the critical rotational velocities for these lower mass stars when accretion stops and they are deposited on the birthline , we have used the pms models of swenson et al . ( 1994 ) and the birthline of palla and stahler ( palla , private communication ) for an accretion rate of @xmath23 m@xmath0/yr to determine the values of m and r on the birthline . the swenson models extend only to 5 m@xmath0 . we have used the models of siess , dufour , & forestini ( 2000 ) to calculate values for stars of 3 - 7 m@xmath0 ; there is good agreement in the overlap region . we have been unable to find calculations of the critical velocity derived from models of pms stars . however , it appears that the 2/3 factor used for high mass stars is a reasonable first approximation even for pms stars . herbst and mundt ( 2005 ) have used the equilibrium shapes of a rotating polytrope of index 1.5 to estimate that the ratio of the polar to equatorial radius for the most rapidly rotating ( @xmath24 days ) low - mass pms stars is @xmath3 0.75 . we have also derived the true rotational velocities for the 5 pms stars in orion with periods less than 0.9 days ( stassun et al . 1999 ) and masses between 0.15 and 0.27 m@xmath0 . if we assume that these stars are rotating at breakup , we find that the lower bound for the factor that should be used to calculate the critical velocity is 0.62 . because these values bracket the factor of 2/3 calculated for high mass stars , we have assumed that this same factor can also be applied to low - mass pms stars . the results are plotted in figure 1 . for low and intermediate mass stars , we have used the values of @xmath1 derived by rhode , herbst & mathieu ( 2001 ) and wsh for stars in the orion association , most of which are less than @xmath25 years old . ( we note that rotation periods have been derived from photometry of many more low mass stars in orion but can not be measured for the o stars , which lack the spot - modulated light curves characteristic of their lower mass brethren . therefore , we choose here to confine our comparison to @xmath1 , which can be measured for stars of all masses . ) . for the high mass stars in our current sample , we use the values of @xmath1 reported in table 1 . we derive masses for the o stars from the evolutionary tracks of schaller et al . ( 1992 ) . in order to do so , we need to know the effective temperature ( @xmath26 ) and bolometric luminosity ( @xmath27 ) . we have used the calibrations given by massey et al . ( 2005 ) to estimate @xmath26 from spectral types given in the papers by massey & thompson ( 1991 ) , hillenbrand et al . 1993 , and massey et al . ( 1995 ) ; given @xmath26 , the bolometric correction can also be obtained from massey et al . the distances to each association have been taken from the papers that provided the spectral types , and intrinsic colors as a function of spectral type given by fitzgerald ( 1970 ) have been used to derive to the reddening to each star . figure 1 shows that nearly all of the stars are rotating at velocities that are less than half the critical velocity on the birthline , and most are rotating at velocities that are less than 30% of the critical velocity . we note that in this figure , we have plotted @xmath17 but have measured @xmath1 . a statistical correction can be made for a group of stars whose inclinations are randomly distributed by multiplying the average @xmath1 by @xmath28 to obtain the average true velocity , which we will call @xmath29 ( chandrasekhar & munch 1950 ; gaige , 1993 ) . if we subdivide the low mass stars into three groups according to mass , we find that the median value of @xmath2 @xmath3 0.14 for the lowest mass group , 0.10 for next lowest mass group , and 0.11 for stars with masses near that of the sun ( see table 2 ) . if we divide the high mass stars into two groups , the medians are 0.13 for stars with masses between 8 and 25 m@xmath0 and 0.20 for the stars with masses greater than 25 m@xmath0 . the discerning reader will note a gap " in the observed data between 3 and 8 m@xmath0 . this gap results from the fact that , while there are many observations of stars near the zams in this mass range , there are very few observations of stars young enough to be near their expected birthline radii . rotation rates for zams stars in this mass range may therefore not be representative of their initial values . rather , such stars are initially deposited on pre - main sequence radiative tracks and will spin up as they contract , resulting in both higher critical velocities and higher rotation rates on the main sequence than would have been observed on the birthline . the results from our limited sample appear to be representative of the results from other studies . stassun et al . ( 1999 ) also find only a few exceptional low mass pms stars in orion with periods close to the critical value , and our results for the low mass orion sample plotted in figure 1 are representative of the range of apparent rotational velocities observed for other groups of pms stars of similar mass . more extensive studies of o stars have found that about 95% of main sequence o stars have values of @xmath1 that are less than 300 km / sec ( penny 1996 ; howarth et al . 1997 ) , with a tail extending to about 400 km / sec . because the effects of limb darkening are severe in stars rotating near the breakup velocity , model calculations for b stars show that line widths may become insensitive to rotation speeds higher than about 90% of the critical velocity for stars viewed equator - on ( townsend et al . only about 5% of the o stars , however , are rotating even as rapidly as half the critical velocity , and so our sample should not be affected by excessive limb darkening . was the rotation of our sample stars also low relative to the critical velocity immediately after accretion stopped ? or has significant angular momentum loss occurred since the stars were deposited on the birthline or , equivalently for the high mass stars , since they first arrived on the zams ? for stars with masses @xmath16 15 m@xmath0 , stellar winds are predicted to carry away angular momentum following deposition on the zams . moreover , as massive stars evolve , the radius increases , thus driving surface rotation speeds toward lower values . the amount of angular momentum lost depends on the age of the star , and unfortunately it is difficult to estimate the ages of massive stars with an accuracy of 1 - 2 myr . the values of t@xmath30 and l@xmath31 derived from the non - lte atmospheric models displace even the youngest o stars to the right of the zams predicted by models of the interior ( schaller et al . 1992 ) by 2000 - 4000 k ( repolust , puls , & herrero 2004 ; martins , schaerer , & hillier 2005 ) . as an alternative method of estimating ages , we note that there is evidence that all of our clusters except cygob2 contain pms stars with masses of 3 - 5 m@xmath0 and temperatures cooler than t@xmath32k . stellar models ( e.g. siess & forestini 2000 ) predict that these stars are less than 2.6 myr old . since the formation of stars in these regions appears to be contemporaneous within 2 - 3 myr with the most massive stars being the youngest ( cf . massey et al . 1995 ) , it is reasonable to assume that the stars in our sample are less than 2.5 myr old . the observations of cyg ob2 do not go faint enough to reach the pms stars , but massey et al . have estimated from the massive stars that the age of cyg ob2 is similar to that of the other associations in our sample . only 4 of the stars in our sample have masses greater than 40 m@xmath0 . quantitative estimates from extant models predict that the surface rotation rate will be reduced by only about 1/3 of the initial value during the first 2.5 myr for stars rotating at 300 km / sec and with masses of 40 m@xmath0 ( meynet & maeder 2000 ) . for most of the stars in our sample , which are of lower mass , the effects of winds and radius changes will be smaller . therefore , we can assume that the currently observed values of @xmath1 for the o stars in our sample differ by not more than 100 km / sec , and typically by much less , from the initial values just after the star was fully assembled . for stars more massive than 40 m@xmath0 , the angular momentum loss is predicted to be larger and , given the uncertainty in ages , we can say little about what their initial angular momentum might have been . some low mass stars must lose angular momentum as they evolve toward the zams along convective tracks ( rebull , wolff & strom , 2004 ; herbst & mundt , 2005 ) . for this reason , the values of @xmath1 for our sample stars could in principle be somewhat lower than their initial values when they were deposited on the birth line . for example , covey et al . ( 2005 ) report that class i / flat spectrum objects rotate on average @xmath3 38 km / sec , or about twice as fast on average as classical t tauri stars . an increase in the average rotation rate of a factor of two would , however , still result in a representative median value of @xmath33 . what we can conclude from the data in figure 1 , therefore , is that throughout the mass range from 0.1 to 50 m@xmath0 the median value of @xmath29 is @xmath3 10 - 20% of the critical velocity ; this difference does not change significantly with mass . a more fundamental quantity than rotation speed is the angular momentum per unit mass . the specific angular momentum of a star is given by j / m = i@xmath34/m . therefore , in order to calculate the specific angular momentum of a star we require values for the moment of inertia ( i ) , the angular rotational velocity , the stellar radius , and the mass . we also assume that the observed surface rotation rate is representative of the rotation of the star as a whole . most published models do not provide moments of inertia , but maeder ( private communication ) has provided us with values for models of stars with masses of 3 , 5 , 9 , and 15 m@xmath0 on the zams . meynet et al . ( 2006 ) have shown that log i scales as log m for massive main sequence stars , and we have used this relationship to extrapolate the values of i from the models we have to the higher mass stars . for the highest mass star for which we have a model ( 15 m@xmath0 ) , the computed tracks predict a change of log i of only 0.11 dex by the time that such a star has completed nearly half its evolution toward core hydrogen exhaustion . since this is a small change , we neglect evolutionary effects in i for the relatively unevolved stars in our sample . the calculated values of j / m are plotted as a function of mass in figure 2 . as seen in this figure , j / m varies slowly with mass over the range 0.2 - 50 m@xmath0 . a fit by eye suggests that over this entire range the upper bound to @xmath35 , which is very close to the value of 0.25 derived by wsh for stars with @xmath6 3 m@xmath0 . the scatter below the upper bound can be attributed to several factors , including differences in viewing angle , loss of angular momentum as low- and intermediate - mass stars evolve down their convective tracks , and depending on the mechanism that determines the initial angular momentum , possibly to differences in stellar properties ( e.g. , magnetic field strength , accretion rate during the stellar assembly phase ) . note that we have included in this plot young stars in the mass range 3 - 8 m@xmath0 for which @xmath1 values are reported by wsh . such objects were excluded from figure 1 on the grounds that these objects are considerably removed from their birthline locations . we include them here because it _ appears _ that stars in this mass range likely conserve angular momentum as they evolve . we do not expect intermediate mass stars to have magnetic fields or strong winds to carry away angular momentum ( see also wsh ) . if this reasoning is correct , then the angular momenta reported here _ may _ accurately reflect their initial values . data for o - type stars have been combined with data in the literature to show that over a mass range of a factor of 250 ( 0.2 - 50 m@xmath0 ) , the specific angular momentum j / m of stars varies slowly and continuously with mass ( @xmath35 ) . nearly all stars in this mass range are rotating at rates that are no more than 30% of the critical velocity calculated from models of stars along the birthline . we conclude that a single mechanism must be at work to keep rotation rates low and at similar values for stars of all masses at birth , this despite the rapid accretion of high angular momentum material from a disk during the stellar assembly phase . in the context of mma and cm models , our results would appear to rule out cm models , which we expect naively to produce more rapid rotation were massive stars formed through mergers . if such mergers take place , they would appear to be the exception rather than the rule . rather , the continuity of angular momentum properties across the whole mass range from m stars to o stars argues for a common formation mechanism to masses as high as 50m@xmath0 and adds one more piece to the mounting evidence ( cf wsh ) that magnetically - mediated accretion through a disk is the main mechanism by which stars of all masses form . with the present data , we can not yet extend this conclusion to still higher masses . we note , however , that penny ( 1996 ) observed several stars with estimated masses larger than 60 m@xmath0 and found none with @xmath1 @xmath10 200 km / sec ( well below the estimated escape velocity of @xmath10 800 km / sec ) . it is unclear whether this means that no massive stars rotate extremely rapidly or whether their strong stellar winds have carried away significant amounts of angular momentum . observations of very young stars with well established ages are needed to determine the angular momentum properties of stars in the range 50 - 100 m@xmath0 . we thank diane harmer of noao , who provided generous assistance both in preparing for our wiyn - hydra observing run and at the telescope , and the referee , georges meynet , for his thorough reading of the manuscript and a number of helpful suggestions . bally , j. , & zinnecker , h. 2005 , aj , 129 , 2281 beuther , h. , & shepherd , d. 2005 , in cores to clusters , in press , astro - ph/0502214 bonnell , i.a . , vine , s.g . & bate , m.r . 2004 , mnras 349 , 735 cesaroni , r. , galli , d. , lodato , g. , walmsley , c. m. , & zhang , q. 2005 , in protostars to planets v , in press chandrasekhar , s. , & munch , g. 1950 , apj , 111 , 142 conti , p. s. , & ebbets , d. , 1977 , apj , 213 , 438 covey , k. r. , greene , t. p. , doppmann , g. w. , & lada , c. j. 2005 , aj , 129 , 276 fitzgerald , m. p. 1970 , a&a , 4 , 234 gaige , y. 1993 , a&a , 269 , 267 hartmann , l. , hewett , r. & calvet , n. 1994 , apj 426 , 669 herbst , w. , & mundt , r. 2005 , apj , 633 , 967 hillenbrand , l. a. , massey , p. , strom , s. e. , & merrill , k. m. 1993 , 106 , 1906 howarth , i. d. , siebert , k. w. , gaitee , a. j. h. , & prinja , r. k. 1997 , mnras , 284 , 265 huang , w. , & gies , d. r. 2006 , apj , submitted johns - krull , c. , valenti , j. & koresko , c. 1999 , apj 516 , 900 krumholz , m. r. 2005 , astro - ph/0511766 mckee , c. f. , & tan , j. c. 2003 , apj 585 , 850 martins , f. , schaerer , d. , & hillier , d. j. 2005 , a&a , 436 , 1049 massey , p. , & thompson , a.b . , 1991 , aj , 101 , 1991 massey , p. , johnson , k. e. , & degioia - eastwood , k. 1995 , aj , 454 , 151 maeder , a. , & meynet , g. 2000 , a&a , 159 , 2000 meynet , g. , & maeder , a. 2000 , a&a , 361 , 101 meynet , g. , ekstrom , s. , maeder , a. , & barblan , f. 2006 , astro - ph/0601339 mihalas , d. 1978 , stellar atmospheres , 2@xmath36 ed . , ( sanfrancisco : freeman ) , p. 256 muzerolle , j. , calvet , n. & hartmann , l. 1998 , apj 492 , 743 . muzerolle , j. , dalessio , p. , calvet , n. & hartmann , l. 2004 , apj 617 , 406 . palla , f. , & stahler , s. w. 1992 , apj , 392 , 667 penny , l. r. 1996 , apj , 463 , 737 . rebull , l. m. , wolff , s. c. , & strom , s. e. 2004 , aj , 127 , 1029 repolust , t. , puls , j. , & herrero , a. 2004 , a&a , 415 , 349 rhode , k. l. , herbst , w. , & mathieu , r. d. 2001 , aj , 122 , 3258 schaller , g. , schaerer , d. , meynet , g. , & maeder , a. 1992 , a&a , 287 , 835 siess , l. , dufour , e. , & forestini , m. 2000 , a&a , 358 , 593 slettebak , a. 1956 , apj , 124 , 173 shu , f. , najita , j. , ostriker , e. , wilkin , f. , ruden , s. & lizano , s. 1994 , apj 429 , 781 . shu , f. , najita , j. , ostriker , e. & shang , h. 1995 , apj 455 , 155 . stassun , k. g. , mathieu , r. d. , mazeh , t. , & vrba , f. j. 1999 , aj , 117 , 2941 swenson , f. j. , faulkner , j. , rogers , f. j. , & iglesias , c. a. 1994 , apj , 425 , 286 townsend , r. h. d. , owocki , s. p. , & howarth , i. d. 2004 , mnras , 350 , 189 wolff , s. c. , strom , s. e. , & hillenbrand , l. a. 2004 , apj , 601 , 979 ic 1805 & & 21 & 2 29 30.47 & + 61 29 44.2 & o9.5vf & 4.49 & 4.94 & 22 & 329 & 17.84 + ic 1805 & bd+60 499 & 118 & 2 32 16.75 & + 61 33 15.0 & 09.5vf & 4.49 & 4.74 & 18 & 50 & 17.01 & + ic 1805 & bd+60 501 & 138 & 2 32 36.28 & + 61 28 25.6 & o7vf & 4.56 & 5.15 & 29 & 179 & 17.77 & + ic 1805 & bd+60 513 & 232 & 2 34 2.53 & + 61 23 10.8 & o8vf & 4.53 & 5.17 & 26 & 304 & 17.87 & + ngc 2244 & hd46056 & 84 & 6 31 20.86 & + 04 50 03.8 & o8vf & 4.53 & 5.02 & 23 & 328 & 17.91 & + ngc 2244 & hd46149 & 114 & 6 31 52.53 & + 05 01 59.2 & o8.5vf & 4.52 & 5.22 & 28 & 72 & 17.24 & + ngc 2244 & hd46150 & 122 & 6 31 55.52 & + 04 56 34.3 & o6e & 4.58 & 5.75 & 52 & 64 & 17.43 & + ngc 2244 & hd46202 & 180 & 6 32 10.47 & + 04 57 59.8 & o9vf & 4.50 & 4.94 & 22 & 42 & 16.97 & + ngc 2244 & & 376 & 6 30 33.33 & + 04 41 27.9 & o9vf & 4.50 & 4.83 & 20 & 54 & 17.08 & + ngc 6611 & & 161 & 18 18 30.95 & -13 43 08.3 & o8.5v & 4.52 & 4.92 & 22 & 118 & 17.46 & + ngc 6611 & & 166 & 18 18 32.22 & -13 48 48.0 & o8.5v & 4.52 & 4.70 & 19 & 83 & 17.33 & + ngc 6611 & & 175 & 18 18 32.73 & -13 45 11.9 & o5.5vf & 4.60 & 5.38 & 37 & 164 & 17.84 & + ngc 6611 & & 188 & 18 18 33.73 & -13 40 58.4 & b0v & 4.48 & 4.43 & 15 & 268 & 17.76 & + ngc 6611 & & 205 & 18 18 36.42 & -13 48 02.4 & o5vf & 4.61 & 5.68 & 52 & 81 & 17.62 & + ngc 6611 & & 280 & 18 18 42.80 & -13 46 50.8 & o9.5v & 4.49 & 4.54 & 17 & 369 & 17.94 & + ngc 6611 & & 314 & 18 18 45.86 & -13 46 30.9 & b0v & 4.48 & 4.80 & 19 & 72 & 17.14 & + ngc 6611 & & 367 & 18 18 52.67 & -13 49 42.6 & o9.5v & 4.49 & 4.59 & 18 & 30 & 16.86 & + ngc 6611 & & 401 & 18 18 56.19 & -13 48 31.0 & o8.5v & 4.52 & 5.07 & 24 & 50 & 17.06 & + ngc 6611 & & 584 & 18 18 23.63 & -13 36 27.9 & o9v & 4.50 & 4.60 & 18 & 18 & 16.66 & + ngc 6611 & hd168504 & & 18 20 34.10 & -13 57 15.7 & o7.5iii & 4.54 & 5.10 & 26 & 84 & 17.37 & + cyg ob2 & 59 & & 20 29 23.62 & + 41 21 41.7 & o8.5v & 4.52 & 5.35 & 34 & 200 & 17.74 & + cyg ob2 & 70 & & 20 29 31.32 & + 41 11 08.8 & o9v & 4.50 & 5.37 & 34 & 47 & 17.07 & + cyg ob2 & 145 & & 20 30 02.56 & + 41 18 14.1 & o9.5v & 4.49 & 4.67 & 18 & 65 & 17.15 & + cyg ob2 & 217 & & 20 30 26.58 & + 41 16 59.1 & o7iiif & 4.56 & 5.50 & 39 & 81 & 17.43 & + cyg ob2 & 227 & & 20 30 29.31 & + 41 15 22.5 & o9v & 4.50 & 4.91 & 21 & 194 & 17.63 & + cyg ob2 & 258 & & 20 30 40.42 & + 41 16 07.3 & o8v & 4.53 & 5.08 & 25 & 174 & 17.65 & + cyg ob2 & 299 & & 20 30 51.26 & + 41 14 58.5 & o7.5v & 4.54 & 5.21 & 30 & 196 & 17.77 & + cyg ob2 & 339 & & 20 31 02.63 & + 41 13 28.8 & o8.5v & 4.52 & 5.03 & 23 & 66 & 17.18 & + cyg ob2 & 376 & & 20 31 11.75 & + 41 14 09.3 & o8v & 4.53 & 4.94 & 23 & 120 & 17.51 & + cyg ob2 & 390 & & 20 31 15.36 & + 41 07 26.0 & o8v & 4.53 & 5.31 & 32 & 101 & 17.45 & + cyg ob2 & 448 & & 20 31 25.51 & + 41 03 10.1 & o6vf & 4.58 & 5.41 & 38 & 171 & 17.83 & + cyg ob2 & 480 & & 20 31 29.81 & + 41 06 50.8 & o7.5v & 4.54 & 5.29 & 33 & 246 & 17.89 & + cyg ob2 & 485 & & 20 31 30.50 & + 41 11 18.6 & o8v & 4.53 & 5.08 & 25 & 112 & 17.46 & + cyg ob2 & 507 & & 20 31 33.35 & + 41 07 21.5 & o8.5v & 4.52 & 4.82 & 20 & 171 & 17.61 & + cyg ob2 & 516 & & 20 31 35.55 & + 40 58 53.2 & o5.5vf & 4.60 & 6.22 & 90 & 80 & 17.64 & + cyg ob2 & 531 & & 20 31 38.38 & + 41 23 09.1 & o8.5v & 4.52 & 5.31 & 32 & 81 & 17.32 & + cyg ob2 & 534 & & 20 31 38.88 & + 41 00 39.7 & o7.5v & 4.54 & 5.18 & 30 & 51 & 17.20 & + cyg ob2 & 555 & & 20 31 43.20 & + 41 25 39.9 & o8v & 4.53 & 5.38 & 34 & 103 & 17.46 & + cyg ob2 & 601 & & 20 31 51.40 & + 41 09 06.2 & o9.5iii & 4.49 & 5.30 & 30 & 77 & 17.22 & + cyg ob2 & 696 & & 20 32 11.66 & + 41 07 14.2 & o9.5v & 4.49 & 5.02 & 23 & 380 & 17.89 & + cyg ob2 & 716 & & 20 32 16.60 & + 40 54 49.7 & o9v & 4.50 & 4.84 & 20 & 132 & 17.46 & + cyg ob2 & 745 & & 20 32 26.13 & + 41 24 40.8 & o7v & 4.56 & 5.22 & 31 & 153 & 17.71 & + cyg ob2 & 771 & & 20 32 41.97 & + 41 21 21.7 & o7v & 4.56 & 5.84 & 55 & 334 & 18.08 & + ngc 7380 & & 31 & 22 47 50.61 & + 58 05 12.2 & o8vf & 4.53 & 4.99 & 23 & 43 & 17.04 & +
we present the results of a study aimed at assessing whether low and high mass stars form similarly . our approach is ( 1 ) to examine the observed projected rotational velocities among a large sample of newly - formed stars spanning a range in mass between 0.2 and 50 m@xmath0 ; and ( 2 ) to search for evidence of a discontinuity in rotational properties that might indicate a difference in the stellar formation process at some characteristic mass . our database includes both recently published values of @xmath1 for young intermediate- and low- mass stars in orion , as well as new observations of o stars located in young clusters and ob associations . we find that the median of the quantity @xmath2 ( observed rotational speed / equatorial breakup velocity ) is typically about 0.15 and shows no evidence of a discontinuity over the full range of stellar masses , while the quantity jsini / m ( derived angular momentum per unit mass ) exhibits a slow , monotonic rise ( j / m @xmath3 @xmath4 ) with increasing mass with no evidence of a discontinuity . we suggest that these observations are most simply interpreted as indicative of a single stellar formation and angular momentum regulation mechanism : one that results in rotation rates well below breakup , and angular momenta per unit mass that differ systematically by no more than a factor of 3 - 4 over a mass range spanning a factor of 250 .
classical current algebras ( also called loop algebras ) and their central extensions , the affine lie algebras , are of fundamental importance in quite a number of fields in mathematics and its applications . these algebras are examples of infinite dimensional lie algebras which are still tractable . they constitute the subclass of kac - moody algebras of untwisted affine type . if one rewrites the original purely algebraic definition in geometric terms the classical current algebras correspond to lie algebra valued meromorphic functions on the riemann sphere ( i.e. on the unique compact riemann surface of genus zero ) which are allowed to have poles only at two fixed points . if this rewriting is done , a very useful generalization ( e.g. needed in string theory ) is to consider the objects over a compact riemann surface of arbitrary genus with more than two points where poles are allowed . the main problem is to introduce a replacement of the grading in the classical case , which is necessary to construct highest weight and fock space representations . this is obtained by an almost - graded structure ( see section [ s : hg ] ) , a weaker structure but still strong enough to do the job . furthermore , to obtain representations of certain types one is forced to pass over to central extensions . such objects ( vector fields , functions , etc . ) and central extensions for higher genus with two possible poles were introduced by krichever and novikov @xcite and generalized by me to the multi - point situation @xcite . these objects are of importance in a global operator approach to conformal field theory @xcite,@xcite . more generally , the current algebra resp . their central extensions , the affine algebras , correspond to symmetries of infinite dimensional systems . their @xmath1-deformed version ( i.e. the quantum affine algebras ) are in close connection with difference equation and with special functions . there is a very interesting direct relation to difference equation . krichever and novikov constructed higher genus analogues of baker akhiezer difference functions which are eigen functions of suitable difference equations . starting from these functions , representations of higher genus affine algebras associated to the lie algebra @xmath2 are obtained . these are related to the two - dimensional toda lattice @xcite . in this write - up of the talk i report on uniqueness and classification results for higher genus multi - point affine lie algebras which i recently obtained . in particular , it turns out that for the current algebra associated to a finite - dimensional simple lie algebra ( e.g. for @xmath3 ) there exists up to equivalence and rescaling a unique non - trivial central extension which extends the almost - grading of the current algebra . the proofs can be found in @xcite . there also further references , historical remarks and corresponding results for the lie algebras of lie algebra valued meromorphic differential operators can be found . the results depend also on a complete classification of central extensions of scalar functions , vector fields and differential operators of krichever novikov type obtained in @xcite . i am indebted to paul terwilliger who asked me to supply explicit examples of such algebras . they can be found in section [ s : zero ] and section [ s : torus ] . let us first consider the nowadays classical affine lie algebras . let @xmath4 be a finite - dimensional complex lie algebra . a special example of fundamental importance is given by the algebra of trace - less matrices @xmath5 with @xmath6:=ab - ba$ ] the commutator as lie product . the current algebra @xmath7 ( sometimes also called loop algebra ) is obtained by tensoring @xmath4 by the ( associative and commutative ) algebra @xmath8}$ ] of laurent polynomials , i.e. @xmath9}$ ] with the lie product @xmath10:=[x , y]\otimes z^{n+m } , \quad x , y\in { \mathfrak{g } } , \ n , m\in{\ensuremath{\mathbb{z}}}.\ ] ] if @xmath4 is a matrix algebra , then @xmath7 can be considered as matrices with laurent polynomials as entries , e.g. a typical element of @xmath11 can be written as @xmath12 where @xmath13 are polynomials in @xmath14 and @xmath15 . by setting @xmath16 the lie algebra @xmath7 is graded ( see ( [ e : classalg ] ) ) . clearly , @xmath7 is an infinite dimensional lie algebra . these algebras are candidates for symmetry algebras of systems with infinitely many independent symmetries . unfortunately , in the process of constructing representations of certain types ( e.g. highest weight representations ) one is forced to `` regularize '' certain natural actions . as a result one obtains only so called _ projective representations _ , which in turn define honest representations of certain central extensions of @xmath7 . what is a central extension @xmath17 of a lie algebra @xmath18 ? as vector space we take @xmath19 . we set @xmath20 and @xmath21 and consider the following product @xmath22:= \widehat{[a , b]}+\psi(a , b)\,t,\quad [ \hat a , t]=0,\ ] ] with @xmath23 a bilinear map . now @xmath17 is a lie algebra if and only if @xmath24 is a lie algebra two - cocycle of @xmath18 ( with values in the trivial module @xmath25 ) . the cocycle conditions are @xmath26,c)+ \psi([b , c],a)+ \psi([c , a],b)=0,\ ] ] for all @xmath27 . two central extensions of the same algebra are called equivalent if they are the same up to some change of basis of the type @xmath28 . in more precise terms , given two extensions defined by @xmath29 and @xmath30 respectively , the two extensions are called equivalent if there exists a linear form @xmath31 such that @xmath32).\ ] ] in other words , the difference is a lie algebra cohomology coboundary . * fact . * the set of equivalence classes of central extensions is via ( [ e : central ] ) in 1 to 1 correspondence to the space of lie algebra two - cohomology classes @xmath33 ( cocycles modulo coboundaries ) . how do we obtain central extensions for our current algebra ? let @xmath34 be an invariant , symmetric bilinear form for @xmath4 . invariance means that @xmath35,c)=\alpha(a,[b , c])$ ] for all @xmath36 . for a simple lie algebra the cartan - killing form is up to a rescaling the only such form . in particular for @xmath3 it is given by @xmath37 . then a central extension @xmath38 is defined by @xmath39=[x , y]\otimes z^{n+m}-\alpha(x , y ) \cdot n\cdot { \delta}_{n}^{-m}\cdot t.\ ] ] to avoid cumbersome notation i dropped the @xmath40 in the notation . it is called the ( classical ) affine lie algebra associated to @xmath4 . by setting @xmath41 ( and using @xmath42 ) the affine algebra is a graded algebra . if the finite dimensional lie algebra @xmath4 is simple then @xmath43 defined via ( [ e : centralclass ] ) is up to equivalence of extensions and rescaling of the central element the only non - trivial central extension of @xmath7 . in this case the algebras are exactly the kac - moody algebras of untwisted affine type @xcite . before we can extend the construction to higher genus we have to geometrize the classical situation . recall that the associative algebra of laurent polynomials @xmath8}$ ] can equivalently be described as the algebra consisting of those meromorphic functions on the riemann sphere ( resp . the complex projective line @xmath44 ) which are holomorphic outside @xmath45 and @xmath46 ( @xmath14 the quasi - global coordinate ) . the current algebra @xmath7 can be interpreted as lie algebra of @xmath4valued meromorphic functions on the riemann sphere with possible poles only at @xmath45 and @xmath46 . the riemann sphere is the unique compact riemann surface of genus zero . from this point of view the next step is to take @xmath47 any compact riemann surface of arbitrary genus @xmath48 and an arbitrary finite set @xmath49 of points where poles of the meromorphic objects will be allowed . in this way we obtain the higher genus ( multi - point ) current algebra as the algebra of @xmath4valued functions on @xmath47 with only possibly poles at @xmath49 . we need gradings , central extensions etc . for this goal we split @xmath49 into two non - empty disjoint subsets @xmath50 and @xmath51 , @xmath52 . in the interpretation of string theory , @xmath50 corresponds to incoming free strings and @xmath51 to outgoing free strings . let @xmath53 be the number of points in @xmath50 . see figure [ figg2 ] for an example given by a riemann surface of genus two with @xmath54 and @xmath55 . let @xmath56 be the associative algebra of functions meromorphic on @xmath47 and holomorphic outside of @xmath49 . in some earlier work @xcite i introduced @xmath57 a certain adapted basis of @xmath56 . for the exact definition i refer to this publication . here we only note that @xmath58 for genus zero and @xmath59 , @xmath60 we get @xmath61 . let @xmath62 be the @xmath53-dimensional subspace of @xmath56 . we have @xmath63 and there exist constants @xmath64 ( independent of @xmath65 and @xmath66 ) such that @xmath67 we call the elements of @xmath68 homogeneous elements of degree @xmath65 . as long as @xmath69 and @xmath70 can not to be chosen to be 0 the algebra is not honestly graded . it is only _ almost - graded_. in a similar way one introduces almost - gradedness for lie algebras . this notion was introduced by krichever and novikov @xcite ( they called it quasi - graded ) and they constructed such an almost - grading in the higher genus and two point case . to find an almost - grading in the multi - point case is more difficult . this weaker grading is enough to introduce and study highest weight representations . as a remark aside : with a special choice of basis one has @xmath71 and the @xmath70 depends in a known manner on the genus @xmath48 and the number of points in @xmath50 and @xmath51 @xcite . the higher genus multi - point current algebra @xmath7 is the tensor product @xmath72 with the lie product @xmath73=[x , y]\otimes ( f\cdot g)\ ] ] and almost - grading @xmath74 the next task is to construct central extensions and to study the question of uniqueness . ( @xcite ) let @xmath34 be an invariant , symmetric bilinear form of @xmath4 and @xmath75 a closed contour on @xmath47 not meeting @xmath49 , then @xmath76 is a lie algebra two - cocycle for the current algebra @xmath7 . hence , it defines a central extension @xmath77 . consequently , there exist central extensions for @xmath7 . but contrary to the classical situation , even if @xmath4 is a simple lie algebra , there will not be a unique nontrivial cocycle class . if we choose essentially different contours @xmath75 the @xmath78 define essentially different central extensions @xmath77 . but in the classical situation we were able to extend our grading of @xmath7 to @xmath43 by assigning a degree to the central element @xmath79 . this will not necessarily be true for all cocycles of the form ( [ e : cocycal ] ) . a 2-cocycle is called _ local _ if there exist @xmath80 and @xmath81 such that @xmath82 given a local cocycle @xmath24 defining a central extension , then by setting @xmath83 ( or any other number ) the almost - grading of @xmath7 extends to the central extension . vice versa , if such an extension of the almost - grading exists , the defining cocycle will be local . we use @xmath84 to denote the subspace of 2-cocycle classes containing a representative which is local . in general , the cocycles @xmath85 are not local . but if we choose as integration contour a smooth contour @xmath86 separating the points in @xmath50 from the points in @xmath51 and which is of winding number 1 , then it can be shown that @xmath87 is a local cocycle . in fact its values can be calculated as @xmath88 we call any such @xmath86 a separating cycle . @xcite let @xmath4 be a finite - dimensional simple lie algebra , @xmath89 its ( higher genus ) current algebra , then @xmath90 and a basis is given by the class of ( [ e : csc ] ) , where @xmath86 is a separating cycle and @xmath91 is a multiple of the cartan killing form . in particular , there exists up to equivalence and rescaling a unique almost - graded non - trivial central extension of @xmath7 . as a side - result we obtain that every local cocycle is cohomologous to a geometric cocycle of the form ( [ e : csc ] ) . the cocycles coming from fock space representations and other type of representations are local . hence we obtain that there exists a unique equivalence class ( up to rescaling of the central element ) of central extensions coming from these representations . in @xcite the more general situation of reductive lie algebras is considered . recall that a finite - dimensional lie algebra @xmath4 is reductive if and only if it is the direct sum ( as lie algebra ) @xmath92 if the abelian summand @xmath93 is missing then @xmath4 is semi - simple . in the semi - simple case it is shown that every local 2-cocycle is cohomologous to a cocycle of the type ( [ e : csc ] ) where @xmath91 is an arbitrary linear combination of the individual cartan killing forms of the summands ( trivially extended to the rest ) . vice versa , such cocycles are local . in particular , we get @xmath94 in the reductive case we have to add another condition . we denote by @xmath95 the lie algebra of meromorphic vector fields on @xmath47 which are holomorphic outside of @xmath49 . a 2-cocycle is called @xmath95-invariant if @xmath96 cocycles of the form ( [ e : cocycal ] ) are @xmath95-invariant . in @xcite it is shown that every @xmath95-invariant local cocycle is cohomologous to ( [ e : csc ] ) . in particular @xmath97 . here the index @xmath95 denotes the classes containing cocycles as representatives which are @xmath95-invariant , and @xmath98 . in the semi - simple case there is no need to pose explicitly @xmath95-invariance as in this case in every local cocycle class there is a unique @xmath95-invariant representative . again , the condition of @xmath95-invariance is automatic if the representations under consideration are representations coming from the larger algebra of lie algebra valued differential operators . the lie algebra of trace - less matrices @xmath3 is simple . hence , the unique non - trivial almost - graded central extension is given ( up to equivalence and rescaling ) by the cocycle @xmath99 the lie algebra of all matrices @xmath100 is the direct sum @xmath101 , where @xmath102 is the abelian summand of scalar matrices . in particular @xmath100 is a reductive lie algebra . following the general results @xmath103 . a basis is given by the elements @xmath29 ( [ e : slcoc ] ) and @xmath104 let us consider the riemann sphere @xmath105 and the set @xmath49 consisting of 3 points . given any triple of 3 points there exists always an analytic automorphism of @xmath106 mapping this triple to @xmath107 , with @xmath108 . in fact @xmath109 would suffice . without restriction we can take @xmath110 due to the symmetry of the situation it is more convenient to take a symmetrized basis of @xmath56 : @xmath111 it is shown in @xcite that it a basis . by more or less direct calculations one can show the structure equation for the current algebra @xmath7 @xmath112 = \begin{cases } [ x , y]\otimes a_{n+m } , & \text{$n$ or $ m$ even } , \\ [ x , y]\otimes a_{n+m}+a^2[x , y]\otimes a_{n+m-2 } , & \text{$n$ and $ m$ odd } , \end{cases}\ ] ] again @xmath109 could be set . the reason to keep @xmath113 is that it can be seen that if we vary @xmath113 over the affine line we obtain for @xmath114 the classical current algebra . in particular , this family gives a deformation . in @xcite it was shown that this deformation is a geometrically non - trivial deformation despite the fact that for @xmath4 simple , @xmath7 is formally rigid , i.e. there does not exists a non - trivial formal deformation . this effect is peculiar to infinite dimensional lie algebras and is discussed in detail there . for the central extension @xmath115 of section [ s : hgcent ] we obtain the defining cocycles ( see @xcite , a.13 and a.14 ) @xmath116 with @xmath117 of course , given a simple lie algebra @xmath4 with generators and structure equations the relations can be written in these terms . for @xmath118 with the standard generators @xmath119 we set @xmath120 , @xmath121 and in the same way @xmath122 and @xmath123 . recall that @xmath124 . we calculate @xmath125&=\begin{cases } h_{n+m } , & \text{$n$ or $ m$ even } \\ h_{n+m}+a^2h_{n+m-2 } , & \text{$n$ and $ m$ odd } \end{cases } \\ [ h_n , e_m]&=\begin{cases } 2e_{n+m } , & \text{$n$ or $ m$ even } \\ 2e_{n+m}+2a^2e_{n+m-2 } , & \text{$n$ and $ m$ odd } \end{cases } \\ [ h_n , f_m]&=\begin{cases } -2f_{n+m } , & \text{$n$ or $ m$ even } \\ -2f_{n+m}-2a^2f_{n+m-2 } , & \text{$n$ and $ m$ odd } \end{cases}\end{aligned}\ ] ] for the central extension we get @xmath126=\begin{cases } h_{n+m}-n\delta_m^{-n } , & \text{$n$ or $ m$ even } \\ h_{n+m}+a^2h_{n+m-2}-n\delta_m^{-n } -a^2(n-1)\delta_m^{-n+2 } , & \text{$n$ and $ m$ odd } \end{cases}\ ] ] the other commutators stay the same . let @xmath127 be a complex one - dimensional torus , i.e. a riemann surface of genus 1 . here @xmath128 denotes the lattice @xmath129 with @xmath130 . the field of meromorphic functions on @xmath131 is generated by the doubly - periodic weierstra @xmath132 function and its derivative @xmath133 fulfilling the differential equation @xmath134 with the @xmath135 pairwise distinct and given by @xmath136 the function @xmath132 is an even meromorphic function with poles of order two at the points of the lattice and holomorphic elsewhere . the function @xmath133 is an odd meromorphic function with poles of order three at the points of the lattice and holomorphic elsewhere . it has zeros of order one at the points @xmath137 and @xmath138 and all its translates under the lattice . @xmath142 = \begin{cases } [ x , y]\otimes a_{n+m } , & \text{$n$ or $ m$ even } , \\ [ x , y]\otimes a_{n+m}+3e_1 [ x , y]\otimes a_{n+m-2 } \\ \ + ( e_1-e_2)(2e_1+e_2 ) [ x , y]\otimes a_{n+m-4 } , & \text{$n$ and $ m$ odd}. \end{cases}\ ] ]
classical affine lie algebras appear e.g. as symmetries of infinite dimensional integrable systems and are related to certain differential equations . they are central extensions of current algebras associated to finite - dimensional lie algebras @xmath0 . in geometric terms these current algebras might be described as lie algebra valued meromorphic functions on the riemann sphere with two possible poles . they carry a natural grading . in this talk the generalization to higher genus compact riemann surfaces and more poles is reviewed . in case that the lie algebra @xmath0 is reductive ( e.g. @xmath0 is simple , semi - simple , abelian , ... ) a complete classification of ( almost- ) graded central extensions is given . in particular , for @xmath0 simple there exists a unique non - trivial ( almost-)graded extension class . the considered algebras are related to difference equations , special functions and play a role in conformal field theory . _ talk presented at the international conference on difference equations , special functions , and applications , munich , july 2005 _
the all - wavelength extended groth strip international survey ( aegis ) is a collaborative effort to obtain both deep imaging covering all major wavebands from x - ray to radio and optical spectroscopy over a large area of sky ( 0.51deg@xmath3 ) with the aim of studying the panchromatic properties of galaxies over the last half of the hubble time . the region studied , the extended groth strip ( egs : @xmath4=14@xmath517@xmath6 , @xmath7=+52@xmath830@xmath9 ) is an extension of and owes its name to a _ hubble space telescope _ ( _ hst _ ) survey consisting of 28 wide - field planetary camera2 ( wfpc2 ) pointings carried out in 1994 by the wfpc team @xcite . this field benefits from low extinction , low galactic and zodiacal infrared emission , and good schedulability by space - based observatories , and has therefore attracted a wide range of deep observations at essentially every accessible wavelength over this comparatively wide field . amongst deep multiwavelength fields , the egs field provides a unique combination of area and depth at almost every waveband observable . it is two ( for _ hst _ ) to four ( for _ spitzer _ and _ chandra _ ) times larger than the combined goods fields @xcite , yet has a similar range of wavelength coverage , making it ideal for studying rare classes of objects that may be absent in smaller fields . the gems field @xcite covers a similar area to similar depths , but was studied by the combo-17 photometric redshift survey rather than a spectroscopic survey . most aegis data sets cover @xmath10.51deg@xmath3 , considerably smaller than the 2deg@xmath3 cosmos field @xcite . however , aegis observations are deeper at most wavelengths , benefiting from greater schedulability and lower backgrounds . spectroscopy of the cosmos field is in progress @xcite , but will not be completed for 35 years . an additional advantage of aegis is that deep _ imaging is available in two bands ( f606w and f814w ) , whereas the cosmos field has been observed in f814w only , while the f850lp imaging in gems is too shallow to study subcomponent colors for most galaxies . even before aegis , the egs region attracted a rich suite of surveys , including both spectroscopy @xcite and panchromatic imaging from both the ground and space , running from x - ray @xcite to ultraviolet ( uv ) and optical @xcite , near infrared ( ir ) @xcite , mid - ir @xcite , submillimeter @xcite , and radio @xcite . aegis has carried this work even further ; for example , the first generation of the deep galaxy redshift survey , deep1 , obtained 620 galaxy redshifts in the wfpc2 groth strip region , now publicly available @xcite . in comparison , the successor deep2 galaxy redshift survey has obtained 9501 redshifts in the egs so far , with thousands more planned . data from deep2 are a linchpin for almost all aegis studies , providing redshifts ; internal kinematics for dynamical masses ; line strengths for star formation rates , agn identification , and gas - phase metallicities ; stellar population ages and metallicites ; etc .. the precision and relatively dense sampling of deep2 redshifts allow for accurate measurement of the local environment of objects in egs , which is a major factor driving galaxy evolution . other surveys at similar redshifts provide weaker environmental measures due to larger redshift errors , lower sampling rates , and/or smaller areas @xcite . furthermore , we can remove cosmic variance fluctuations from observed aegis abundances by comparing redshift distributions to the other three , widely - separated deep2 fields . ten instrument teams and a number of theorists are now collaborating on aegis nearly 100 scientists in half a dozen countries . the first fruits of this collaboration are presented in this issue of _ astrophysical journal letters_. these papers make use of the power of the combined aegis dataset in a variety of ways . six _ letters _ investigate the nature of rare objects found in the aegis field , illustrating the benefits of covering an area wide enough to find them in @xcite . these studies take advantage of the full multiwavelength coverage from aegis , which provides each object s spectral energy distribution ( sed ) in detail ; we also explore the range of seds exhibited by a wider set of galaxies in @xcite . five more _ letters _ investigate the drivers and evolution of star formation in galaxies using the wide array of indicators available from this multifaceted dataset , based on uv , ir , and radio continuum as well as optical emission lines @xcite . three _ letters _ focus on the optical properties of active galactic nuclei at @[email protected] identified using both deep x - ray and ir data , and explore their relationship to their large - scale structure environment @xcite . two _ letters _ test for evolution in the relationships between mass measures to @xmath0@xmath11 : between stellar mass and gas kinematics within individual galaxies @xcite , and between x - ray gas emission and galaxy kinematics in groups of galaxies @xcite . the remaining two _ letters _ use deep2 spectroscopy to investigate the nature of objects which are extremely red in optical - ir color , finding that a substantial fraction of the population lies at @[email protected] @xcite . these papers present only the first results from the aegis survey ; we are just beginning to reach the potential of this manyfaceted dataset . this _ letter _ gives details on the aegis data which have been obtained so far ; we only describe those survey data sets in the egs field which are used in this special issue here . large portions of this dataset , including the deep2 spectra and _ hst_/acs imaging , will be publicly released in 2007 , making it a legacy for the entire community . the basic parameters of the aegis multiwavelength data sets are listed in table[tab : data_tbl ] and their sky coverage is shown in figure[fig : skymap ] . the acquisition and reduction of the aegis data sets and the derivation of source photometry catalogs [ from which panchromatic spectral energy distributions ( seds ) are measured ] are described in detail below . _ spitzer _ irs mid - ir spectra exist for a relatively small number of galaxies ; the details of these observations are in @xcite and @xcite . cccccccc /acis & hb & [email protected]@xmath11 & 3.1 ( 4kev ) & 8.2@xmath1210@xmath13ergs@xmath14@xmath15 & 3200 & 0.67 & 200 + & sb & [email protected]@xmath11 & 12.4 ( 1kev ) & 1.1@xmath1210@xmath13ergs@xmath14@xmath15 & 2500 & 0.67 & 200 + galex & fuv & 5.5@xmath11 & 1539 & 25 ( ab ) [ @xmath16 & 8720 & 1.13 & 58 + & nuv & 5.5@xmath11 & 2316 & 25 ( ab ) [ @xmath16 & 2.35@xmath1210@xmath17 & 1.13 & 120 + cfht / megacam & @xmath18 & @xmath101.1@xmath11 & 3700 & @xmath127 ( ab ) & @xmath1910@xmath20 & 1 & 6.1 + ( cfhls/_current _ ) & @xmath21 & @xmath101.0@xmath11 & 4850 & 28.3 ( ab ) & @xmath1910@xmath20 & 1 & 6.5 + & @xmath22 & @xmath100.9@xmath11 & 6250 & @xmath127.5 ( ab ) & @xmath1910@xmath20 & 1 & 15 + & @xmath23 & @xmath100.9@xmath11 & 7700 & @xmath127 ( ab ) & @xmath1910@xmath20 & 1 & 47 + & @xmath24 & @xmath100.9@xmath11 & 8850 & 26.4 ( ab ) & @xmath1910@xmath20 & 1 & 3.6 + cfht / cfh12k & @xmath25 & 1@xmath11 & 4389 & 24.5 ( ab ) [ 8@xmath26 & 1.45@xmath1210@xmath20 & 1.31 & 6.5 + & @xmath27 & 1@xmath11 & 6601 & 24.2 ( ab ) [ 8@xmath26 & 1.45@xmath1210@xmath20 & 1.31 & 3.6 + & @xmath28 & 1@xmath11 & 8133 & 23.5 ( ab ) [ 8@xmath26 & 1.45@xmath1210@xmath20 & 1.31 & 4.7 + _ hst_/acs & f606w ( @xmath29 ) & 0.1@xmath11 & 5913 & 28.75 ( ab ) [ 5@xmath26 & 4.0@xmath1210@xmath20 & 0.197 & 2.3 + ( wfc ) & f814w ( @xmath28 ) & 0.1@xmath11 & 8330 & 28.10 ( ab ) & 3.9@xmath1210@xmath20 & 0.197 & 2.1 + _ hst_/nicmos & f110w ( @xmath30 ) & 0.35@xmath11 & 1.10@xmath2 m & 25.7 ( ab ) [ 10@xmath26 & 3.3@xmath1210@xmath20 & 0.0128 & 2.6 + ( nic3 ) & f160w ( @xmath31 ) & 0.35@xmath11 & 1.59@xmath2 m & 25.5 ( ab ) [ 10@xmath26 & 3.3@xmath1210@xmath20 & 0.0128 & 2.6 + palomar / wirc & @xmath30 & 1@xmath11 & 1.25@xmath2 m & 23 ( vega ) & 7.64@xmath1210@xmath17 & 0.2 & 18 + & @xmath32 & 1@xmath11 & 2.14@xmath2 m & 20.6 ( vega ) & 5.37@xmath1210@xmath17 & 0.7 & 11 + _ spitzer_/irac & band1 & 1.8@xmath11 & 3.6@xmath2 m & 0.9@xmath2jy & 1.66@xmath1210@xmath20 & 0.33 & 10.1 + & band2 & 2.0@xmath11 & 4.5@xmath2 m & 0.9@xmath2jy & 1.68@xmath1210@xmath20 & 0.33 & 10.1 + & band3 & 2.2@xmath11 & 5.8@xmath2 m & 6.3@xmath2jy & 4.90@xmath1210@xmath17 & 0.33 & 10.1 + & band4 & 2.2@xmath11 & 8.0@xmath2 m & 5.8@xmath2jy & 4.86@xmath1210@xmath17 & 0.33 & 10.1 + _ spitzer_/mips & 24@xmath2 m & 5.9@xmath11 & 23.7@xmath2 m & 77@xmath2jy & 1.76@xmath1210@xmath17 & 0.534 & 1.68 + & 70@xmath2 m & 19@xmath11 & 71.4@xmath2 m & 10.3mjy & 1275 & 0.498 & 0.84 + vla & 6 cm & 1.2@xmath11 & 6 cm & 0.55mjybeam@xmath14 [ 10@xmath26 & 88.9 & 5.2 & 0.5 + & 20 cm & 4.2@xmath11 & 20 cm & 100@xmath2jybeam@xmath14 & 1075 & 0.64 & 400 + telescope / instr . & . range & spec . res . & spatial psf & lim . mag . & area & num . time + ( mode ) & ( ) & ( ) & ( fwhm ) & & ( deg@xmath3 ) & & ( ks ) + keck / deimos & 64009100 & 1.4 & 0.6@xmath111@xmath11 & @xmath33=24.1 & 0.5 & 17,600 & 3.6 + ( deep2/spectra ) & & & & & ( _ final _ ) & ( _ final _ ) & [ tab : data_tbl ] the egs region has been surveyed at x - ray wavelengths by _ chandra _ using the advanced ccd imaging spectrometer ( acis ) . the observations consist of eight individual pointings obtained with the acis - i instrument with a pixel scale of @xmath190.49@xmath11 and a field of view of @xmath1917@xmath9@xmath1217@xmath9 . the total exposure time per pointing is @xmath1200ks split into at least four separate integrations obtained at different epochs . the data reduction is performed using the ciao data analysis software version3.2 . after identifying hot pixels and cosmic - ray afterglows , the raw data are processed through the ciao acis_process_events task , which applies the latest gain map and charge transfer inefficiency corrections . the observations are then screened using the standard asca grade set ( event grades 0 , 2 , 3 , 4 , and 6 ) and inspected for flaring spikes , when the background deviates by more than 3@xmath34 from the mean quiescent value . bad time intervals amount to @xmath35% of the total exposure time per pointing . individual observations of a given _ chandra _ pointing are then merged into a single event file . the final mosaic images are used to detect sources in a number of energy bands , including 0.52.0kev , 2.07.0kev , and 0.57.0kev . the source detection is based on pre - selection of candidate sources using the ciao wavdetect task with a low likelihood probability threshold ( @xmath36 ) , followed by aperture count extraction using the 70% psf radius and a local background determination to estimate the source significance ( for details see * ? ? ? the final catalog comprises sources with poisson false probability @xmath104@xmath1210@xmath37 . point - source fluxes are estimated by integrating the net counts within the 90% encircled energy radius at the position of the source . we adopt a power - law sed with @xmath38=1.4 and galactic neutral hydrogen column density @xmath39=1.3@xmath1210@xmath40@xmath15 , appropriate for the egs field . the limiting fluxes in standard bands are listed in table[tab : data_tbl ] . only a subset of the aegis _ chandra _ observations has been analysed at this point and therefore used for followup analysis ( see * ? ? ? * ) . at the time of writing the aegis _ chandra _ observations represent the third deepest x - ray survey in the sky . the hubble deep field north survey is @xmath15@xmath12 deeper over a @xmath15@xmath12 smaller area @xcite . the _ chandra _ deep field south @xcite covers about half the aegis survey area mostly to a similar depth , with a smaller central region that is twice as deep . the vast majority of x - ray sources detected in the egs field are active galactic nuclei ( agn ) . at the target redshift of the deep2 survey , @xmath0=1 , the aegis limiting luminosity of 7@xmath1210@xmath41ergs@xmath14 corresponds to @xmath10.005@xmath12@xmath42 , where @xmath42 is the characteristic agn luminosity at this redshift @xcite . the galaxy evolution explorer ( galex ) images a 1.25@xmath8 diameter field . galex images of the egs are built from stacks of 42 far - ultraviolet ( fuv ) and 87 near - ultraviolet ( nuv ) separate one - orbit images obtained in 2003 , 2004 , and 2005 and processed using version 4.1 of the galex pipeline . the combined exposure times are 5.8@xmath1210@xmath43s for fuv and 1.2@xmath1210@xmath44s for nuv . the raw photon count images are flat - fielded and calibrated using relative response maps . the resulting calibrated intensity images are in units of photons@xmath14 , where 1photons@xmath14 corresponds to 18.82 and 20.08 ab mag for the fuv and nuv bands , respectively . the background in the images is estimated using poissonbg , a program written for galex data that uses poisson rather than gaussian statistics to clip suspected sources from the background map . finally , the source catalogs are derived from the background - subtracted images with sextractor @xcite . more details about the galex pipeline can be found at http://www.galex.caltech.edu . the egs is one of four 1deg@xmath3 fields covered by the ongoing canada - france - hawaii telescope legacy survey ( cfhtls ) deep survey . this field ( labelled `` d3 '' by cfhtls ) has been observed for a total of 114hr using the megacam imager on the 4 m canada - france - hawaii telescope ( cfht ; * ? ? ? * ) from april 2003 to the present ; a total integration time of 330hr is planned by the end of the five - year survey . this time is divided amongst five broad - band filters : @xmath18 , @xmath21 , @xmath22 , @xmath23 , and @xmath24 . only images with seeing fwhm smaller than [email protected]@xmath11 ( depending on the band ) are included in the survey . five - sigma point source detection limiting ab magnitudes in the current data set range from 26.4 ( in @xmath24 ) to 28.3 ( in @xmath21 ) . the catalogs used in aegis papers are based on a subset of the cfhtls data set with total exposure times of 1.7/1.8/4.1/13.0/1.0hr in @xmath18/@xmath21/@xmath22/@xmath23/@xmath24 . after visual rejection of defective exposures , elixir - processed frames were run through the astrogwyn and photgwyn software packages , improving both astrometric and photometric calibrations significantly @xcite ; they were then corrected for distortions and coadded using swarp . photometry was then obtained using the double image mode of sextractor @xcite , with @xmath23 used as reference image . the principal cfhtls measurements used here are kron aperture ( sextractor mag_auto ) ab magnitudes . in addition , photometric redshifts for cfhtls sources were determined by applying the gwynz code to photometry within matched , 1@xmath11 radius apertures @xcite . additional details of the procedures used are available at : http://www.astro.uvic.ca/grads/gwyn/cfhtls/d3.html . the egs was imaged in @xmath25 , @xmath27 , and @xmath28 bands using the cfh12k mosaic camera @xcite on cfht . this 12k@xmath128k mosaic camera has a scale of 0.21@xmath11pixel@xmath14 and a field of view of 0.70@[email protected]@xmath8 ( with the longer axis oriented east - west ) . the @xmath27-band seeing ranged from 0.75@xmath111@xmath11 fwhm in the four separate cfht12k pointings covering the egs ; integration times were @xmath11hr in @xmath25 and @xmath27 and @xmath12hr in @xmath28 . the @xmath27-band magnitudes were measured within circular apertures of radius 3@xmath45 , where @xmath45 is the @xmath34 of a gaussian fit to the image profile ; for objects where 3@xmath45@xmath101 , a 1@xmath11 radius aperture was used instead . the @xmath25@xmath46@xmath27 and @xmath27@xmath46@xmath28 colors of each object were measured using a 1@xmath47 radius aperture . the resulting @xmath48 photometry was calibrated to the ab system within the native cfht12k passbands ( which differ significantly from the kron - cousins system , particularly in @xmath28 ) using stars observed by the sloan digital sky survey ( sdss ; * ? ? ? * ) ; the @xmath48 stellar locus is used to ensure consistency of the color system between cfh12k pointings . details of the data reduction , astrometry , star - galaxy separation , and catalog construction can be found in @xcite ; the resulting catalogs are available at http://deep.berkeley.edu/dr1 . deep _ hst _ images of the egs were obtained with acs as part of go program 10134 ( pi : m. davis ) . the egs was imaged in the @xmath29 ( f606w ) and @xmath28 ( f814w ) bands during the period 2004 june to 2005 march . a mosaic pattern consisting of 21@xmath123=63 contiguous `` tiles '' was used to cover an effective area of @xmath110.1@[email protected]@xmath9=710.9arcmin@xmath3 following the irac imaging strip ( [ sec : irac ] ) . the exposure times per `` tile '' were 2260 and 2100s in the @xmath29 and @xmath28 bands , respectively . tiles were observed at a position angle of 130@xmath8 , or rotated by multiples of 90@xmath8 relative to this value to meet guide star constraints . each tile was observed in a 4-pointing dither pattern in each filter , in order to achieve half - pixel dithering at the center of acs wfc , bridge the detector gap , and improve tile overlap . the final mosaic is gap - free and each pixel is observed at least three times . dithered pointings were combined with the stsdas multidrizzle package using a square kernel . the final images have a pixel scale of 0.03@xmath11pixel@xmath14 with a psf of 0.12@xmath11 fwhm . the 5@xmath34 limiting magnitudes for a point source are @xmath49 ( ab ) and @xmath50 ( ab ) within a circular aperture of radius 0.12@xmath11 ( @xmath150pixel area ) . for an extended object , the 5@xmath34 limiting magnitudes are @xmath51 ( ab ) and @xmath52 ( ab ) for a circular aperture of radius @xmath53 ( @xmath1314pixel area ) . we detected objects in summed acs @xmath29+@xmath28 images and constructed initial galaxy segmentation maps using the sextractor galaxy photometry software @xcite and a detection threshold of 1.5@xmath34 and 50pixels . these detection maps and the acs zeropoints @xcite were applied to each band separately to create the acs photometric catalogs . the egs is one of the four fields observed by the deep2 collaboration @xcite . here we briefly describe the deep2 data in egs ; for more details see @xcite for maskmaking algorithms , faber et al . ( in prep . ) for full survey details , and cooper et al . ( in prep . ) for data reduction pipelines . targets were selected for deep2 spectroscopy from the cfht12k @xmath48 imaging described in [ sec : cfh12k ] . eligible deep2 targets have 18.5@xmath54@[email protected] , @xmath5520% probability of being a galaxy ( based on angular size , @xmath25@xmath46@xmath27/@xmath27@xmath46@xmath28 colors , and @xmath27 mag ) , and surface brightness brighter than : @xmath56 where @xmath57 is the area of the aperture used to measure the cfht12k @xmath27 magnitude ( [ sec : cfh12k ] ) ; all magnitudes are ab . each object is given a weight based on its probability of being a galaxy , its @xmath27 magnitude , and whether or not meets the deep2 color cut used to eliminate low-@xmath0 objects in other fields [ galaxies with ( @xmath25@xmath46@xmath27)@xmath102.45(@xmath27@xmath46@xmath28)@xmath460.5 , ( @xmath27@xmath46@xmath28)@xmath551.1 , or ( @xmath25@xmath46@xmath27)@xmath100.5 all pass this cut ] . this weight is used when randomly selecting amongst multiple objects that can not be observed simultaneously due to deimos slitmask constraints @xcite . fainter objects ( particularly those with @[email protected] and expected @[email protected] from the color cut ) are given lower weight in order to sample a range of luminosities and roughly equal numbers of galaxies below and above @[email protected] ( faber et al . in prep . ) . selection probabilities for each potential target are known to @xmath101% ; the median is @xmath5570% for objects with @[email protected] , falls to 54% at @[email protected] , and is flat at 73% for @[email protected] . all spectra were taken with the deimos spectrograph @xcite at the keckii telescope . each observation uses a unique aluminum mask milled with @xmath1150 1@xmath11 wide and @xmath553@xmath11 long slitlets over a 16@xmath9@xmath124@xmath9 area , which is observed for a minimum of 1hr ( until a target signal - to - noise is reached ) divided amongst 3 or more sub - exposures . the 1200linemm@xmath14 grating used yields a dispersion of @xmath10.33pixel@xmath14 and a spectral resolution of fwhm=1.4 . the typical wavelength range of the spectra is @xmath165009100 , varying modestly with slit position . slitlets are tilted up to 30@xmath8 to follow the photometric major axes of extended targets . all deep2 spectra were reduced with an idl pipeline heavily modified from the idlspec2d package designed for sdss ( burles & schlegel in prep . ) . spectra were extracted using both boxcar and optimized gaussian weighting . the pipeline determines a set of candidate redshifts for each object by fitting a linear combination of templates at each possible @xmath0 and finding local minima in @xmath59 ; these redshifts are then evaluated and selected amongst by deep2 team members using a graphical interface . two significant features must match the templates for a secure redshift ( quality @xmath60=3 or 4 ) ; a resolved [ oii ] 3727 doublet is counted as two features . based on both repeated observations and tests with multiwavelength photometry , we estimate that @xmath615% of @xmath60=3 redshifts ( obtained for 11% of egs targets ) are incorrect , while @xmath610.5% of highest - confidence , @xmath60=4 redshifts ( 60% of egs targets ) are incorrect . lower quality redshifts are considered insecure or ambiguous and not used for any analyses . objects with repeated observations ( 310 out of the 13570 galaxies observed so far in egs ) have an rms redshift uncertainty of 30kms@xmath14 . secure redshifts have been obtained for 9501 galaxies in egs , with median redshift 0.74 . objects at @[email protected] tend not to have strong features in the deep2 spectral window ; such objects appear to comprise the bulk of deep2 redshift failures ( c. steidel , priv . comm . ) . deep _ hst _ near infrared camera and multi - object spectrometer ( nicmos ) nic3 images of the egs were obtained as `` parallels '' of the acs images ( [ sec : acs ] ) as part of go program10134 ( pi : m.davis ) . each nic3 field was observed in the @xmath30 ( f110w ) and @xmath31 ( f160w ) bands for 2560s per band ( one _ hst _ orbit ) . the field of view of a single nic3 pointing is 51.2@[email protected]@xmath11 . the resulting 63nic3 fields cover a combined area of @xmath1946arcmin@xmath3 . out of the 63 , 58 fully overlap with the acs imaging mosaic ; the remaining five nic3 pointings coincide with other aegis data sets . since the nic3 psf is undersampled , we developed a 4-point dither pattern that simultaneously provides optimal sub - pixel dithering for acs wfc and nic3 to improve the final resolution of the reduced images . the nicmos nic3 images were processed in similar way as the acs images . basic image reductions were performed with the stsdas.calnica routine : flatfielding and corrections for dark current , bias , variable quadrant bias , amplifier glow , cosmic ray persistence , detector nonlinearity , pixel defects , bad imaging regions , cosmic rays of unusual size , and the count - rate dependent nonlinearity . the four pointings per tile per filter were combined with the stsdas.multidrizzle package using a square kernel . the final images have a scale of 0.1@xmath11pixel@xmath14 with a psf of 0.35@xmath11 fwhm . sources were detected using summed nic3 @xmath30+@xmath31 images using sextractor @xcite , using a detection threshold of 1@xmath34 and minimum size of 10pixels . photometry was performed on the individual images using circular apertures of diameter 0.52@xmath11 calibrated using the zero - points in the nicmos data handbook . near - ir observations of the egs in the @xmath30 and @xmath62 bands were obtained using the wide - field infrared camera ( wirc ) on the palomar 5 m telescope . the observations were carried out between 2003 and 2005 . wirc has an effective field of view of 8.1@[email protected]@xmath9 , with a scale of 0.25@xmath11pixel@xmath14 . the egs observations consist of 33 wirc overlapping pointings in @xmath62 and 10 pointings in @xmath30 , each with 4@xmath1230s exposures dithered over a non - repeating 7@xmath11 pattern . typical total exposure times per band at any given location within the egs are 12hr . the net seeing fwhm ranges from 0.8@xmath11 to 1.2@xmath11 . photometric calibration was carried out by referencing standard stars during photometric conditions . the final images were made by combining individual mosaics obtained over several nights . the images were processed using a double - pass reduction pipeline developed specifically for wirc . the total area covered in the @xmath62 band is 2400arcmin@xmath3=0.67deg@xmath3 , with about a third of this area covered in the @xmath30 band . the _ spitzer _ mid - ir observations were carried out as part of guaranteed time observing ( gto ) program number 8 , using time contributed by g. fazio , g. rieke , and e. wright . the infrared array camera ( irac ; * ? ? ? * ) observations were performed in two epochs , 2003 december and 2004 june / july . each irac exposure covered a 5.12@[email protected]@xmath9 field of view with 256@xmath12256 pixels and a scale of 1.2@xmath11pixel@xmath14 . at each of 52 positions in a 2@xmath8@xmath1210@xmath9 map there were dithered 200s exposures at 3.6 , 4.5 , and 5.8@xmath2 m , together with 208 dithered 50s exposures taken concurrently at 8.0@xmath2 m . data processing began with the basic calibrated data produced by version11 of the spitzer science center irac pipeline . individual frames were corrected for the ` muxbleed ' artifact near bright stars . mosaicing was done using custom idl scripts : each individual frame was distortion - corrected and projected onto a reference frame , and the frames were combined by averaging with 3@xmath34-clipping . rejection of cosmic rays , scattered light , and other image artifacts was accomplished by the sigma - clipping during mosaicing and also facilitated by having the observations done at two position angles differing by @xmath1180@xmath8 . the scale of the mosaics , 0.6@xmath11pixel@xmath14 , sub - samples the native irac pixel scale by a factor of two . the two shorter - wavelength irac bands are more sensitive than the longer - wavelength bands . to make catalogs , sources in the irac mosaics were identified using daophot / find , and photometered in a 3@xmath11 diameter aperture . aperture corrections to the irac calibration photometry aperture of 12.2@xmath11 ( multiplicative factors of 2.07 , 2.15 , 2.45 , and 2.68 in the four bands ) were applied . for each object , neighboring objects in a 200-pixel box were subtracted before photometry . both 3.6@xmath2m- and 8.0@xmath2m - selected catalogs were generated : photometry for the objects in each was centered on the position in the mosaic in the selected band . there are about 73,000 objects in the 3.6@xmath2m - selected catalog , many of which are undetected in the 8.0@xmath2 m mosaic ; the 8.0@xmath2 m catalog contains only 16,000 objects . _ spitzer _ far - ir observations with the mips instrument @xcite were also carried out as part of the same gto program . data were obtained in january and june 2004 using the slow rate mips scan mode with legs 2.4@xmath8 long . the mips 24@xmath2 m channel has a 5.4@[email protected]@xmath9 field of view ( 128@xmath12128 array of 2.55@xmath11 pixels ) . the 70@xmath2 m channel has a 5.2@[email protected]@xmath9 field of view ( 32@xmath1232 array of 9.98@xmath11 pixels ) , but only half of the array is functional . the 160@xmath2 m channel has a 5.3@[email protected]@xmath9 field of view ( 20@xmath123 array of 16@xmath11@xmath1218@xmath11 pixels ) , but one row of the array is not operational . the final mosaic covers an area @xmath12.4@xmath8@xmath1210@xmath9 . the effective integration time at 24@xmath2 m is @xmath11500s for locations near the long centerline of the strip , decreasing to @xmath1700s 5@xmath9 from the centerline . at 70 and 160@xmath2 m , the average integration times are @xmath1700 and 100spixel@xmath14 respectively . the data were reduced and mosaiced with the mips data analysis tool @xcite . sources were identified and photometry extracted with psf fitting using the daophot software @xcite . radio continuum observations at 6 cm ( 4.8ghz , c band ) were obtained at the very large array ( vla ) in bna configuration for a total of 19hr during 2003 october 1113 ( the continuum mode maximizes sensitivity but decreases the effective field of view ) . at 4.8ghz the vla antennas have a primary beam fhwm of 9@xmath9 . the mapping grid contained 74 pointings , spaced 5@xmath9 apart , providing roughly uniform sensitivity over a 17@xmath9@xmath122@xmath8 strip . each pointing center was observed for 15min . phase stability was sufficient to yield astrometric accuracy better than 0.1@xmath11 rms , and flux density was calibrated relative to 3c286 . images containing bright ( @xmath5510mjybeam@xmath14 ) point sources were self - calibrated . the data were reduced using the aips software package . for each of the 74 pointings , we created a 2048@xmath122048pixel image with a scale of 0.4@xmath11pixel@xmath14 . to avoid clean bias we clean - ed the images to a flux level of 260@xmath2jybeam@xmath14 , corresponding to @xmath14@xmath34 rms , which took @xmath1100200 iterations . for each quarter of the length of the strip , overlapping images were combined into a mosaic using the linmos task in the miriad software package . the rms noise in the mosaiced images is 42@xmath2jybeam@xmath14 . source candidates were extracted using sextractor @xcite and photometry was performed with jmfit in aips on the individual pointing images , with corrections for delay beam distortions . altogether 51 radio components ( some of which may be double radio sources ) were detected at @xmath6310@xmath34 significance . further details are in table 1 , and the source list and identification with irac and deep2 counterparts are given by @xcite . observations at 20 cm ( 1.4ghz , l band ) are described by @xcite where the resulting catalog , aegis20 , containing 1123 discrete radio emitters is published in electronic form . briefly , spectral - line data were obtained using the vla in b configuration , with correlator mode ` 4 ' , for a total of 110hr in 2003 december and 2005 april june . eight overlapping pointings were observed spanning the length of the egs , concentrating on the six furthest from the bright source , 3c295 . for each pointing the @xmath130@xmath9 fwhm primary beam was blanketed with 37 images , each comprising 512@xmath12512pixels with a scale of 0.8@xmath11pixel@xmath14 , with 1020 additional images centered on more distant , bright sources ( including 3c295 ) . central images of the six main pointings were mosaiced , correcting for the primary beam response and excluding data beyond the half - power point . the resulting image covers 0.04 , 0.36 , and 0.64deg@xmath3 to @xmath64 limits of 50 , 75 , and 100@xmath2jybeam@xmath14 . source detection for aegis20 followed that described by @xcite . this _ letter _ has described the multiwavelength dataset in the extended groth strip assembled by the aegis collaboration . the remaining _ letters _ in this special issue discuss a wide range of scientific results derived from these data . by combining deep observations at almost every wavelength available over @xmath65 deg@xmath3 of sky with the relatively high resolution spectroscopy and dense sampling of the deep2 galaxy redshift survey , aegis is making possible many unique studies of the evolution of galaxies over more than half the history of the universe . the high - quality internal kinematics and environment measurements in egs are unmatched amongst deep multiwavelength fields . the dataset is continuing to grow and its potential is only beginning to be tapped ; we expect aegis to provide a legacy long into the future . this research has made use of nasa s astrophysics data system bibliographic services . the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community . we are most fortunate to have the opportunity to conduct observations from this mountain . alc and jan are supported by nasa through hubble fellowship grants hf-01182 and hf-01165 awarded by stsci , which is operated by aura , inc . , for nasa , under contract nas 5 - 26555 . jml acknowledges support from the noao leo goldberg fellowship , nasa / stsci grants go-10134 and ar-10675 , nasa nag5 - 11513 grant to p. madau , and a calspace grant to d. c. koo . lam s work was carried out at jpl / caltech , under a contract with nasa . sak would like to thank eddie bergeron for assistance with reducing and calibrating the nicmos data . alexander , d. m. , et al . 2003 , , 126 , 539 barger , a. j. , cowie , l. l. , mushotzky , r. f. , yang , y. , wang , w .- h . , steffen , a. t. , & capak , p. , , 129 , 578 beck - winchatz , b. , & anderson , s. f. 1999 , , 117 , 2582 bertin , e. , & arnouts , s. 1996 , , 117 , 393 biggs , a. d. , & ivison , r. j. 2006 , , submitted boulade , o. , et al . 2003 . , spie , 4841 , 72 brunner , r. j. , connolly , a. j. , & szalay , a. s. 1999 , , 516 , 563 cardiel , n. , elbaz , d. , schiavon , r. p. , willmer , c. n. a. , koo , d. c. , phillips , a. c. , & gallego , j. 2003 , , 584 , 76 coil , a. l. , et al . 2004 , , 617 , 765 conselice , c. j. , et al . 2006 , , in press ( this issue ; astro - ph/0607242 ) cooper , m. c. , et al . 2005 , , 634 , 833 coppin , k. , halpern , m. , scott , d. , borys , c. , & chapman , s. 2005 , , 357 , 1022 crist ' obal - hornillos , d. , balcells , m. , prieto , m. , guzm ' an , r. , gallego , j. , cardiel , n. , serrano , ' a. , & pell ' o , r. 2003 , , 595 , 71 cuillandre , j .- c . , luppino , g. , starr , b. , & isani , s. 2001 , in proceedings of semaine de lastrophysique franaise , eds . f. combes , d. barret , & f. thvenin ( les ulis : edp - sciences ) , 605 davis , m. , et al . 2003 , spie , 4834 , 161 davis , m. , gerke , b. f. , newman , j. a. , & the deep2 team 2005 , asp conf . 339 : observing dark energy , 339 , 128 faber , s. m. , et al . 2003 , spie , 4841 , 1657 fang , t. , et al . 2006 , , submitted ( this issue ) fazio , g. g. , et al . 2004 , , 154 , 10 flores , h. , et al . 1999 , , 517 , 148 fomalont , e. b. , et al . 1991 , , 102 , 1258 georgakakis , a. , et al . 2006 , , in press ( this issue ; astro - ph/0607274 ) gerke , b. f. , et al . 2006 , , submitted ( this issue ) giaconni , r. , et al . 2002 , , 139 , 369 giavalisco , m. , et al . 2004 , , 600 , l93 gordon , k. d. , et al . 2005 , , 117 , 503 gwyn , s. , et al . 2006 , in preparation hines , d. c. , et al . 2006 , , 638 , 1070 hopkins , a. m. , connolly , a. j. , & szalay , a. s. 2000 , , 120 , 2843 houck , j. r. , et al . 2004 , , 154 , 18 huang , j .- s . , et al . 2006 , , submitted ( this issue ) ivison , r. , et al . 2006 , , in press ( this issue ; astro - ph/0607271 ) kassin , s. a. , et al . 2006 , , submitted ( this issue ) kirby , e. n. , et al . 2006 , , submitted ( this issue ) koekemoer , a. m. , & scoville , n. z. 2005 , new astron . , 49 , 461 konidaris , n. , et al . 2006 , , submitted ( this issue ) le floch , e. , et al . 2006 , , submitted ( this issue ) lehmer , b. d. , et al . 2005 , , 161 , 21 lilly , s. j. , hammer , f. , le fevre , o. , & crampton , d. 1995 , , 455 , 75 lilly , s. , & the zcosmos team . 2005 , the messenger , 121 , 42 lin , l. , et al . 2006 , , submitted ( this issue ; astro - ph/0607272 ) magnier , e. a. , & cuillandre , j .- c . 2004 , , 116 , 449 miyaji , t. , et al . 2004 , , 127 , 3180 moustakas , l. , et al . 2006 , , submitted ( this issue ; astro - ph/0607239 ) nandra , k. , et al . 2005 , , 356 , 568 nandra , k. , et al . 2006 , , in press ( this issue ; astro - ph/0607270 ) noeske , k. g. , et al . 2006a , , submitted ( this issue ) noeske , k. g. , et al . 2006b , , submitted ( this issue ) pierce , c. m. , et al . 2006 , , submitted ( this issue ) rhodes , j. , refregier , a. , & groth , e. j. 2000 , , 536 , 79 rieke , g. h. , et al . 2004 , , 154 , 25 rix , h .- w . , et al . 2004 , , 152 , 163 sarajedini , v. , et al . 2006 , , in press ( astro - ph/0605370 ) simard , l. , et al . 2002 , , 142 , 1 sirianni , m. , et al . 2005 , , 117 , 1049 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2003 , , 592 , 728 stetson , p. b. 1987 , , 99 , 191 symeonidis , m. , et al . 2006 , , submitted ( this issue ) vogt , n. p. , et al . 2005 , , 159 , 41 weiner , b. j. , et al . 2005 , , 620 , 595 weiner , b. j. , et al . 2006 , , submitted ( this issue ) willner , s. p. , et al . 2006 , , submitted wilson , g. , et al . 2006 , , submitted ( this issue ) york , d. g. , et al . 2000 , , 120 , 1579
in this the first of a series of _ letters _ , we present a description of the panchromatic data sets that have been acquired in the extended groth strip region of the sky . our survey , the all - wavelength extended groth strip international survey ( aegis ) , is intended to study the physical properties and evolutionary processes of galaxies at @xmath0@xmath11 . it includes the following deep , wide - field imaging data sets : _ chandra_/acis x - ray ( 0.510kev ) , galex ultraviolet ( 12002500 ) , cfht / megacam legacy survey optical ( 36009000 ) , cfht / cfh12k optical ( 45009000 ) , _ hubble space telescope_/acs optical ( 44008500 ) , palomar / wirc near - infrared ( 1.22.2@xmath2 m ) , _ spitzer_/irac mid - infrared ( 3.68.0@xmath2 m ) , _ spitzer_/mips far - infrared ( 2470@xmath2 m ) , and vla radio continuum ( 620 cm ) . in addition , this region of the sky has been targeted for extensive spectroscopy using the deimos spectrograph on the keck ii 10 m telescope . our survey is compared to other large multiwavelength surveys in terms of depth and sky coverage .
estimating the direction of arrival ( doa ) of multiple signals impinging on an array of sensors from observation of a finite number of array snapshots has been extensively studied in the literature @xcite . maximum likelihood estimators ( mle ) and cramr - rao bounds ( crb ) , derived under the assumption of additive white gaussian noise , and either for the so - called conditional or unconditional model @xcite , serve as references to which newly developed doa estimators have been systematically compared . in many instances however , additive noise is usually colored and , consequently , the problem of doa estimation in spatially correlated noise fields has been studied , see e.g. , @xcite . when the spatial covariance matrix of this additive noise is known a priori , maximum likelihood estimators and cramr - rao bounds are changing in a straightforward way with whitening operations . the new statistical problem appears when the covariance matrix of the additive noise is not known a priori and information about this matrix is substituted by a number of independent and identically distributed ( i.i.d . ) training samples , that form the so - called secondary training sample data set . in many cases one can assume that the statistical properties of the training noise data are the same as per noise data within the primary training set data : such conditions are usually referred to as the supervised training conditions . therefore , under these conditions , one has two sets of measurements , one primary set @xmath2 which contains signals of interest ( soi ) and noise , and a second set @xmath3 ( secondary training set ) which contains noise only . examples of this problem formulation are numerous in the area of passive location and direction finding . for instance , in the so - called over - sampled 2d hf antenna arrays , ionospherically propagated external noise is spatially non white @xcite , and some parts of hf spectrum ( distress signals for example ) with no signals may be used for external noise sampling @xcite . despite its relevance in many practical situations , this problem has been relatively scarcely studied @xcite . for parametric description of the gaussian noise covariance matrix with @xmath4 the unknown parameter vector , in @xcite , the authors derive the cramr - rao bound for joint soi parameters ( doa ) @xmath5 and noise parameters @xmath4 estimation , assuming a conventional unconditional model , i.e. , @xmath6 and @xmath7 where @xmath8 stands for the complex gaussian distribution whose respective parameters are the mean , row covariance matrix and column covariance matrix . @xmath9 is the usual steering matrix with @xmath10 the vector of unknowns doa , @xmath11 denotes the waveforms covariance matrix and @xmath12 corresponds to the noise covariance matrix , which is parameterized by vector @xmath4 . in many cases however , the gaussian assumption for the predominant part of the noise can not be advocated . typical example is the hf external noise , heavily dominated by powerful lighting strikes @xcite . evidence of deviations from the gaussian assumption has been demonstrated numerous times for different applications , with the relevance of the compound - gaussian ( cg ) models being justified @xcite . in essence , the individual @xmath13-variate snapshot of such a noise over the face of an antenna array may be treated as a gaussian random vector , whose power can randomly fluctuate from sample to sample . cg models belong to a larger class of distributions , namely multivariate elliptically contoured distributions ( ecd ) @xcite . for the sake of clarity , we briefly review the main definitions of ecd . a vector @xmath14 follows an ec distribution if it admits the following stochastic representation @xmath15 where @xmath16 means `` has the same distribution as '' . in , @xmath17 is a non - negative real random variable and is independent of the complex random vector @xmath18 which is uniformly distributed over the complex sphere @xmath19 . the matrix @xmath20 is such that @xmath21 where @xmath22 is the so - called scatter matrix , and we assume here that @xmath22 is non - singular . the probability density function ( p.d.f . ) of @xmath23 can then be written as @xmath24 where @xmath25 stands for proportional to . the function @xmath26 is called the density generator and satisfies finite moment condition @xmath27 . it is related to the p.d.f . of the modular variate @xmath17 by @xmath28 . going back to our scenario of two data sets @xmath29 and @xmath30 , we assume that they are independent , and that their columns are independent and distributed ( i.i.d . ) according to . in other words , one has @xmath31 and @xmath32 , where @xmath33 and @xmath34 are i.i.d . variables drawn from @xmath35 , and @xmath36 and @xmath37 are i.i.d . random vectors uniformly distributed on the unit sphere . it then follows that the joint distribution of @xmath38 is given by @xmath39 where @xmath40 and [ p(xp , xs ) ] @xmath41 where @xmath42 . additionally , we assume that @xmath43 depends on a parameter vector @xmath5 while @xmath22 depends on @xmath4 . our objective is then to estimate @xmath44 from @xmath45 . let us emphasize an essential difference of the problem in with respect to the typical problem of target detection in cg clutter @xcite . there , within each range resolution cell the clutter is perfectly gaussian and therefore the optimum space - time processing is the same as per the standard gaussian problem formulation . it is the data dependent threshold and clutter covariance matrix ( in adaptive formulation ) that needs to be calculated from the secondary data , if not known a priori @xcite . in the problem , the soi doa estimation should be performed on a number @xmath46 of ecd i.i.d . primary training samples , and maximum likelihood doa estimation algorithm and crb should be expected to be very different from the gaussian case . the paper is organized in the following way . in section [ section : crb ] , we derive a general expression of the fim for elliptically distributed noise using two data sets . section [ section : doak ] focuses on the case of doa estimation in @xmath0-distributed noise . in section [ section : crbk ] , we derive conditions under which the fim is bounded / unbounded , and provide a sufficient condition for unboundedness of the fim with general elliptical distribution . the maximum likelihood estimate , as well as an approximation , are derived in section [ section : mlek ] . in the same section , we derive lower and upper bounds on the mean - square error of the mle for non - regular estimation conditions , i.e. , when the fisher information matrix is unbounded . numerical simulations serve to evaluate the performance of the estimators in section [ section : numerical ] and our conclusions are drawn in section [ section : conclu ] . in this section , we derive the crb for estimation of parameter vector @xmath47 from the distribution in . the fisher information matrix ( fim ) for the problem at hand can be written as @xcite @xmath48 where we used the fact that @xmath49 hence , the total fim is the sum of two matrices @xmath50 , with straightforward definition from . in order to derive each matrix , we will make use of the general expression of the fisher information matrix for ecd recently derived in @xcite . first , let us introduce @xmath51 where @xmath52 . then , we have from @xcite that the @xmath53-th element of the fisher information matrices is given by @xmath54 { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{j}}\ } } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{k}}\ } } \nonumber \\ & + \frac{\alpha_{2 } { t_{p}}}{m(m+1 ) } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{j}}{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{k}}\}}\end{aligned}\ ] ] @xmath55 { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{j}}\ } } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{k}}\ } } \nonumber \\ & + \frac{\alpha_{2 } { t_{s}}}{m(m+1 ) } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{j}}{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}_{k}}\}}\end{aligned}\ ] ] where @xmath56 . since @xmath22 depends only on @xmath4 , it follows that @xmath57 takes the following form @xmath58 with @xmath59 { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{j}}\ } } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{j}}\ } } \nonumber \\ & + \frac{\alpha_{2 } { t_{s}}}{m(m+1 ) } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{j}}{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{k}}\}}\end{aligned}\ ] ] where @xmath60 . let us now consider @xmath61 . using the fact that @xmath22 depends only on @xmath4 and @xmath62 depends only on @xmath5 , @xmath61 is block - diagonal , i.e. , @xmath63 with @xmath64 { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{j}}\ } } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{j}}\ } } \nonumber \\ & + \frac{\alpha_{2 } { t_{p}}}{m(m+1 ) } { { \mathrm{tr}}\{{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{j}}{{{\mathbf{r}}}^{-1}}{{{\mathbf{r}}}^n_{k}}\}}.\end{aligned}\ ] ] the whole fim is thus given by @xmath65 the crb for estimation of @xmath5 is obtained as the upper - left block of the inverse of the fim and is thus simply @xmath66 . similarly to the gaussian case , the crb for estimation of @xmath5 in the conditional model is the same as if @xmath22 was known . as for the crb for estimation of @xmath4 , it is the same as if we had a set of @xmath67 noise only samples . we address the specific problem where the primary data can be written as @xmath68 where @xmath69 follows a gamma distribution with shape parameter @xmath70 and scale parameter @xmath71 , i.e. , its p.d.f . is given by @xmath72 which we denote as @xmath73 , and @xmath74 . the noise component is known to follow a @xmath0 distribution and @xmath75 in admits a ces representation similar to with @xmath76 . the p.d.f . of @xmath33 in this case is given by @xmath77 where @xmath78 is the modified bessel function . note that the @xmath79-th order moment of @xmath80 is @xmath81 where we used the fact @xcite that @xmath82 the density generator is thus here @xmath83 where , for the sake of notational convenience , we have dropped the subscript @xmath84 . the fim for @xmath0-distributed noise can be obtained from the fim for gaussian distributed noise and the calculation of the scalar @xmath85^{2}\right\}}\ ] ] for @xmath86 . for the signal parameters part only , we indeed have @xmath87 where the subscript @xmath88 and @xmath89 stand for @xmath0-distributed and gaussian distributed noise . using the fact that @xmath90 , it follows that @xmath91 \nonumber \\ & = - { q}^{\frac{\nu - m-1}{2 } } \beta^{-1/2 } k_{m+1-\nu}\left(2 \sqrt{{q}/ \beta}\right).\end{aligned}\ ] ] it then ensues that @xmath92 and thus @xmath93^{2 } \right\ } } \nonumber \\ & = \frac{2 \beta^{-\frac{\nu+m}{2}-1}}{\gamma(\nu ) \gamma(m ) } \int_{0}^{\infty } { q}^{\mu-2 + \frac{\nu+m}{2 } } \frac{k^{2}_{m+1-\nu}\left(2 \sqrt{{q}/ \beta}\right)}{k_{m-\nu}\left(2 \sqrt{{q}/ \beta}\right ) } d { q}\nonumber \\ & = \frac{\beta^{\mu-2}}{2^{2\mu+\nu+m-4}\gamma(\nu ) \gamma(m ) } \int_{0}^{\infty } z^{2\mu+\nu+m-3 } \frac{k^{2}_{m+1-\nu}\left(z\right)}{k_{m-\nu}\left(z\right ) } d z.\end{aligned}\ ] ] a formula for the fim in case of @xmath0-distributed noise was derived in @xcite based on the compound gaussian representation . while it resembles our derivations based on the fim for ecd derived in @xcite , it does not match exactly our expression herein . moreover , we study herein the _ existence of the fim _ and derive _ a closed - form approximation of the fim_. let us investigate the conditions under which the integral @xmath94 converges . towards this end , let us use the following inequality which holds for @xmath95 and @xmath96 @xcite @xmath97 it follows that @xmath98 the first integral converges for @xmath99 while the second converges for @xmath100 . hence , for @xmath101 , one has @xmath102 accordingly , one has , for @xmath96 @xmath103 which implies that @xmath104 the first integral converges for @xmath101 and the second converges for @xmath105 . in the former case , one has @xmath106 consequently , we conclude that _ the integral converges only for _ @xmath101 : for @xmath107 this implies that @xmath108 which is verified . in contrast , when @xmath109 , one must have @xmath110 . in other words , the term in _ the fim corresponding to the noise parameters is always bounded _ since it depends on @xmath111 only . the situation is different for signal parameters . in an unconditional model where @xmath22 would depend on signal parameters as well , the fim is bounded . in contrast , in the conditional model where signal parameters are embedded in the mean of the distribution , the fim corresponding to signal parameters _ is bounded only for @xmath110 _ : otherwise , it is unbounded . the latter case corresponds to the so - called non regular case corresponding to distributions with singularities , as studied e.g. , in @xcite . before pursuing our study of the fim for the specific case of @xmath0-distributed noise , let us make an important observation . for the @xmath0 distribution , we have just proven that @xmath112 does not exist for @xmath113 . however , see , @xmath114 exists if and only if @xmath115 and @xmath116 . the latter condition implies that , when @xmath113 , @xmath117 does not exist . observe that convergence of the latter integral is problematic in a neighborhood of @xmath118 , since for @xmath119 , @xmath120 as @xmath121 is a density . therefore , at least for @xmath0-distributed noise , if @xmath122 does not exist , then @xmath123 is unbounded . at this stage , one may wonder if this property extends to any other elliptical distribution . it turns out that this is indeed the case , as stated and proved in the next proposition . whatever the p.d.f . of the modular variate @xmath33 , if @xmath124 then @xmath125 . for the sake of notational convenience , we temporarily omit the subscript @xmath84 and use @xmath17 instead of @xmath33 . let us first observe that @xmath126 since @xmath127 , one can write @xmath128 which implies that @xmath129^{2}.\end{aligned}\ ] ] therefore @xmath130^{2 } p({q } ) d{q}\nonumber \\ & = ( m-1)^{2 } \int_{a}^{b } { q}^{-1 } p({q } ) d{q}- 2 ( m-1 ) \int_{a}^{b } p'({q } ) d{q}\nonumber \\ & + \int_{a}^{b } { q}\left [ \frac{\partial \ln p({q})}{\partial { q } } \right]^{2 } p({q } ) d{q}\nonumber \\ & = ( m-1)^{2 } \int_{a}^{b } { q}^{-1 } p({q } ) d{q}- 2 ( m-1 ) \left [ p(b ) - p(a ) \right ] \nonumber \\ & + \int_{a}^{b } { q}\left [ \frac{\partial \ln p({q})}{\partial { q } } \right]^{2 } p({q } ) d{q}. \end{aligned}\ ] ] the third term of the sum is always positive . in the second term , we have that @xmath131 . it follows that divergence of @xmath132 is a sufficient condition for divergence of @xmath133 . as said before @xmath134 exists , and therefore a sufficient condition for @xmath135 to be undounded is that @xmath136 is unbounded . let us now go back to the @xmath0-distributed case and investigate whether it is possible to derive a simple expression for @xmath137 and subsequently @xmath138 , assuming that @xmath101 . towards this end , let us make use of @xmath139 to write that @xmath140 the last term is obviously not possible to obtain in closed - form so that we use a `` large @xmath141 '' approximation of the modified bessel function @xcite @xmath142 which results in @xmath143 therefore , @xmath144 we finally have @xmath145 if the large @xmath141 approximation is made from the start , then one has @xmath146 so that @xmath147 and hence @xmath148 figure [ fig : alphamu_approx ] compares the approximations in and , as well as a method which uses random number generation to approximate @xmath138 based on its initial definition in . more precisely , we generated a large number of random variables @xmath149 and replace the statistical expectation of by an average over the so - generated random variables . as can be observed from figure [ fig : alphamu_approx ] , the @xmath150 approximations provide very close values , which enable one to validate the closed - form expressions in and . in and . @xmath151 and @xmath152.,width=283 ] we now focus on maximum likelihood ( ml ) estimation of direction of arrival @xmath153 , signal waveforms @xmath154 and covariance matrix @xmath22 in the model @xmath155 where @xmath156 , and @xmath157 . the joint distribution of @xmath158 is given by @xmath159^{\frac{\nu - m}{2 } } k_{m-\nu}\left ( 2 \sqrt{{{\mathbf{y}}_{{t_{s}}}}^{h } { { { \mathbf{r}}}^{-1}}{{\mathbf{y}}_{{t_{s}}}}/ \beta } \right ) \nonumber \\ & \times \prod_{{t_{p}}=1}^{{t_{p } } } \left [ { { { \mathbf{z}}}_{{t_{p}}}}^{h } { { { \mathbf{r}}}^{-1}}{{{\mathbf{z}}}_{{t_{p}}}}\right]^{\frac{\nu - m}{2 } } k_{m-\nu}\left ( 2 \sqrt { { { { \mathbf{z}}}_{{t_{p}}}}^{h } { { { \mathbf{r}}}^{-1}}{{{\mathbf{z}}}_{{t_{p}}}}/ \beta } \right)\end{aligned}\ ] ] where @xmath160 . joint estimation of all parameters appears to be very complicated and hence we will proceed in two steps . at first , we assume that @xmath22 is known and derive the ml estimates of @xmath161 and @xmath154 . then , @xmath22 is substituted for some estimate obtained from observation of @xmath162 only . assuming that @xmath22 is known , one needs to maximize with respect to @xmath161 and @xmath154 @xmath163^{h } { { { \mathbf{r}}}^{-1}}\left[{{\mathbf{x}}_{{t_{p}}}}- { { \mathbf{a}}}(\phi ) s_{{t_{p}}}\right ] \right)\ ] ] where @xmath164 is given by . since @xmath164 is monotonically decreasing , see , it follows that @xmath165 is maximized when the argument of @xmath164 is minimized . however , @xmath166^{h } { { { \mathbf{r}}}^{-1}}\left[{{\mathbf{x}}_{{t_{p}}}}- { { \mathbf{a}}}(\phi ) s_{{t_{p}}}\right ] \nonumber \\ & = \left[{{\mathbf{a}}}^{h}(\phi ) { { { \mathbf{r}}}^{-1}}{{\mathbf{a}}}(\phi ) \right ] \left| s_{{t_{p } } } - \frac{{{\mathbf{a}}}^{h}(\phi ) { { { \mathbf{r}}}^{-1}}{{\mathbf{x}}_{{t_{p}}}}}{{{\mathbf{a}}}^{h}(\phi ) { { { \mathbf{r}}}^{-1}}{{\mathbf{a}}}(\phi ) } \right|^{2 } \nonumber \\ & + { { \mathbf{x}}_{{t_{p}}}}^{h } { { { \mathbf{r}}}^{-1}}{{\mathbf{x}}_{{t_{p}}}}- \frac{\left| { { \mathbf{a}}}^{h}(\phi ) { { { \mathbf{r}}}^{-1}}{{\mathbf{x}}_{{t_{p}}}}\right|^{2}}{{{\mathbf{a}}}^{h}(\phi ) { { { \mathbf{r}}}^{-1}}{{\mathbf{a}}}(\phi)}.\end{aligned}\ ] ] therefore , for any @xmath161 , @xmath165 is maximized when @xmath167 it ensues that one needs now to maximize , with respect to @xmath161 @xmath168 with @xmath169 . in order to avoid calculation of a modified bessel function and thus in order to simplify estimation , we propose to make use of the `` large @xmath141 '' approximation of the modified bessel function given in to write @xmath170 this approximation results in an approximate maximum likelihood ( aml ) estimator of @xmath161 which consists in maximizing @xmath171^{\nu - m}.\ ] ] note that @xmath172\ ] ] which should be compared to the concentrated log likelihood function in the gaussian case , as given by @xmath173.\ ] ] a few remarks are in order about these estimates , in particular about the behavior of the aml estimator in the case of unbounded fim , i.e. , when @xmath174 . first , note that all estimates will be a function of @xmath175 where @xmath176 is the projection onto the orthogonal complement of @xmath177 . compared to , the logarithm operation in will strongly emphasize those snapshots @xmath75 for which @xmath178 is small . let us thus investigate the properties of this statistic , when evaluated at the _ true _ value of signal doa @xmath153 . using the fact that @xmath179 , where @xmath180 and @xmath181 is a short - hand notation for @xmath182 , one has @xmath183 for small @xmath70 ( @xmath174 ) , it follows that , in the vicinity of @xmath153 , the snapshot with minimal @xmath178 is more or less the snapshot for which @xmath69 is minimum , hence the snapshot for which noise power is minimum , which makes sense . if we let @xmath184 , then its cumulative density function ( c.d.f . ) is given by @xmath185 } & = 1 - \left ( 1 - { { \mathrm{pr}}\left[\tau_{{t_{p } } } \leq \eta\right ] } \right)^{{t_{p } } } \nonumber \\ & = 1 - \left [ 1 - \gamma\left(\nu , \eta \beta^{-1}\right ) \right]^{{t_{p } } } \end{aligned}\ ] ] which is shown in figure [ fig : cdf_min_tau_t=4 ] . obviously , with small @xmath70 , the snapshot which corresponds to the minimum value of @xmath69 exhibits a very high signal to noise ratio and , due to the emphasizing effect of the @xmath186 operation in , the performance of the aml estimator is likely to be driven mainly by this particular snapshot . this is illustrated in figure [ fig : mse_rknown_order_tau_vs_t_vs_nu ] where we display the mean - square error ( mse ) of the aml estimate which uses all @xmath46 snapshots and the mse of an hypothetical aml estimator which would use only the snapshot @xmath187 corresponding to the minimum value of @xmath69 . the scenario of this simulation is described in the next section . this figure shows a marginal loss of the aml estimator using @xmath187 only , as compared to the full aml estimator , especially for small @xmath70 . . @xmath188 and @xmath189.,width=283 ] . @xmath22 known , @xmath188 and @xmath190db.,width=283 ] let us thus analyze the behavior of the aml estimators . for the sake of notational convenience , let @xmath191 and @xmath192 denote the aml estimator using @xmath46 snapshots with @xmath0-distributed noise and the aml estimator using the snapshot @xmath187 corresponding to the minimal @xmath69 , respectively . observe that , when using a single snapshot @xmath187 , minimizing is equivalent to minimizing the gaussian likelihood function in with @xmath193 . since @xmath187 exhibits a high signal to noise ratio , @xmath192 is close to @xmath153 , one can make a taylor expansion and relate the error @xmath194 to the error @xmath195 as @xmath196 where @xmath197 is some vector that depends essentially on the derivatives of @xmath198 @xcite and whose expression is not needed here . one can simply notice that @xmath197 would be the same with gaussian noise and a single snapshot , since maximizing or is equivalent when one snapshot is used . this implies that @xmath199 observe that @xmath200 is the mean - square error ( mse ) that would obtained in gaussian noise and a single snapshot , which is about @xmath46 times the mse obtained in the gaussian case and using @xmath46 snapshots , and the latter is approximately the gaussian crb . the mse of @xmath192 depends on @xmath201 where @xmath202 is the minimum value of a set of @xmath46 independent and identically distributed ( actually gamma distributed ) variables . therefore , in order to obtain @xmath201 , one must consider statistics of extreme values , a field that has received considerable attention for a long time , see e.g. , @xcite . it turns out that only asymptotic ( as @xmath203 ) results are available and we build upon them to derive the rate of convergence of @xmath204 . first , note that @xmath205 } & = { { \mathrm{pr}}\left[u_{{t_{p } } } \geq { t_{p}}^{-1/ \nu } x\right ] } \nonumber \\ & = \left ( { { \mathrm{pr}}\left[\tau_{{t_{p } } } \geq { t_{p}}^{-1/ \nu } x\right ] } \right)^{{t_{p } } } \nonumber \\ & = \left(1 - { { \mathrm{pr}}\left[\tau_{{t_{p } } } \leq { t_{p}}^{-1/ \nu } x\right ] } \right)^{{t_{p } } } \nonumber \\ & = \left [ 1 - \gamma\left(\nu,\beta^{-1}{t_{p}}^{-1/ \nu } x\right ) \right]^{{t_{p}}}.\end{aligned}\ ] ] now since @xmath70 is small and @xmath46 is large , @xmath206 is very small and we can approximate @xmath207^{-1 } y^{a}$ ] , which yields @xmath208 } & = \left [ 1 - \gamma\left(\nu,\beta^{-1}{t_{p}}^{-1/ \nu } x\right ) \right]^{{t_{p } } } \nonumber \\ & \simeq \left [ 1 - \frac{\beta^{-\nu } { t_{p}}^{-1 } x^{\nu}}{\nu \gamma(\nu ) } \right]^{{t_{p } } } \nonumber \\ & \simeq { \exp \left\{-\frac{\beta^{-\nu } { t_{p}}^{-1 } x^{\nu}}{\nu \gamma(\nu)}\right\}}^{{t_{p } } } \nonumber \\ & = { \exp \left\{-\frac{\beta^{-\nu } x^{\nu}}{\nu \gamma(\nu)}\right\}}.\end{aligned}\ ] ] it follows that asymptotically , @xmath209 converges to the distribution in , whose probability density function is @xmath210 using integration by parts , it follows that @xmath211^{\infty}_{0 } + \int_{0}^{\infty } { \exp \left\{-\frac{\beta^{-\nu } x^{\nu}}{\nu \gamma(\nu)}\right\ } } dx \nonumber \\ & = \beta \nu^{1/\nu -1 } \gamma(\nu)^{1/\nu } \int_{0}^{\infty } z^{1/\nu -1 } { \exp \left\{-z\right\ } } dz \nonumber \\ & = \beta \nu^{1/\nu -1 } \gamma(\nu)^{1/\nu } \gamma(\nu^{-1 } ) \triangleq c(\nu,\beta).\end{aligned}\ ] ] one can then conclude that , as @xmath46 goes to infinity , @xmath212 c(\nu,\beta ) { t_{p}}^{-1/\nu}.\ ] ] therefore , in the case of @xmath174 , the mse of @xmath192 decreases as @xmath213 , a rate of convergence much faster than the usual @xmath214 . note that this case corresponds to unbounded fim . such rates of convergence are also found with distributions possessing singularities @xcite . as for the aml estimate obtained from @xmath46 snapshots , namely @xmath191 , its mse is upper - bounded by that @xmath192 ( since it uses all snapshots , including @xmath187 ) , and is lower - bounded by the mse that would be obtained if @xmath215 for @xmath216 , and this mse is @xmath214 times the mse of @xmath192 . additionally , as said before , we have @xmath217 where @xmath218 is the gaussian crb using @xmath46 snapshots . hence , one can bound the mse of @xmath191 as @xmath219 as will be illustrated in the next section , the upper bound is rather tight , while the lower bound is much lower than the actual mse . when @xmath22 is not known , then the secondary data @xmath162 can be used to estimate it . the maximum likelihood estimator is obtained ( for @xmath220 ) as the solution ( up to a scaling factor ) to the following implicit equation @xcite @xmath221 @xmath222 can be obtained through an iterative procedure , whose convergence is guaranteed under the assumptions made @xcite . in order to avoid evaluation of the modified bessel function , one can use the large @xmath141 approximation of @xmath223 in to define @xmath224 as the solution to the fixed - point solution @xmath225 note that @xmath224 is more or less the well - known tyler fixed - point estimator @xcite , which again can be obtained from an iterative procedure whose convergence is guaranteed @xcite . the drawbacks of the two above estimators are that 1)they are suited to a @xmath0 distribution for the noise and 2)@xmath226 is required to be larger than @xmath13 . in order to gain robustness against these problems , a solution is to use normalized data @xmath227 whose distribution is independent of that of the noise , and to use regularization . more precisely , we suggest to resort to the following scheme @xcite [ fp_r(eta ) ] @xmath228^{-1 } { { { \mathbf{z}}}_{{t_{s } } } } } + \eta { { \mathbf{i}}}_{m } \\ { \hat{{{\mathbf{r}}}}}_{k+1}(\eta ) & = \frac{m}{{{\mathrm{tr}}\{\breve{{{\mathbf{r}}}}_{k+1}(\eta)\}}}\breve{{{\mathbf{r}}}}_{k+1}(\eta).\end{aligned}\ ] ] and define @xmath229 since convergence of this iterative scheme has been proved @xcite . the very good performance of this scheme has been illustrated in various applications , see e.g. , @xcite , where discussions on how to select the regularization parameter @xmath230 can also be found . we assume a linear array of @xmath188 elements spaced a half - wavelength apart and we consider the simple scenario of a single source impinging from @xmath231 embedded in unit power @xmath0-distributed noise . the covariance matrix @xmath22 is given by @xmath232 with @xmath233 . the exact and approximate maximum likelihood estimators , which consists in maximizing @xmath234 in @xmath235 in were implemented using the matlab function ` fminbnd ` , and the maximum was searched in the interval @xmath236 $ ] where @xmath237 is the half - power beamwidth of the array . the signal waveform was generated from i.i.d . gaussian variables with power @xmath238 and the signal to noise ratio ( snr ) is defined as @xmath239 . the asymptotic gaussian crb , multiplied by the scalar @xmath240 was used as the bound for @xmath0-distributed noise . for the regularized covariance matrix estimator @xmath241 of , the value of @xmath230 was set to @xmath242 . @xmath243 monte - carlo simulations were used to evaluate the mean - square error ( mse ) of the estimates . in figures [ fig : mse_vs_t_ts=32_nu>1 ] and [ fig : mse_vs_t_ts=32_nu<1 ] we plot the crb ( for @xmath110 ) or the lower and upper bounds of when @xmath244 , as well as the mse of the ml and aml estimators , as a function of @xmath46 , and compare the case where @xmath22 is known to the case where it is estimated from with @xmath245 snapshots in the secondary data . the following observations can be made : * there is almost no difference between the mle and the amle , and therefore the latter should be favored since it does not require evaluating modified bessel functions . * the mse in the case where @xmath22 is known is lower than that when @xmath22 is to be estimated , which is expected . however , the difference is smaller when @xmath244 : in other words , it seems that adaptive whitening is not so much penalizing with small @xmath70 while it seems more crucial for @xmath246 . indeed , for small @xmath70 , what matters most is the fact that some snapshots are nearly noiseless , and this is more influential than obtaining a very good whitening . * the decrease of the mse for @xmath110 is roughly of the order @xmath214 . when @xmath244 , this rate is significantly increased and the mse decreases very quickly as @xmath206 , as predicted by the analysis above . this rate of convergence is also observed in figure [ fig : mse_multi_rknown_vs_t_vs_nu ] where we consider a scenario with two sources at @xmath247 . * the upper bound in seems to provide quite a good approximation of the actual mse , at least for @xmath46 large enough . the influence of @xmath226 is investigated in figure [ fig : mse_vs_ts_t=16 ] , where one can observe that about @xmath248 is necessary for the performance with estimated @xmath22 to be very close to the performance for known @xmath22 . however , as indicated above , this is less pronounced when @xmath244 , where the difference becomes smaller with lower @xmath226 . finally , we investigate whether the rate of convergence of the mle or amle when @xmath70 varies is impacted by a small amount of gaussian noise . more precisely , we run simulations where the data is generated as @xmath249 where @xmath250 , i.e. , the noise is a mixture of @xmath0-distributed noise and gaussian distributed noise . the covariance matrix of the noise is now @xmath251 and we use the aml estimator assuming that the noise has a @xmath0 distribution with parameter @xmath70 and known covariance matrix @xmath252 . in figure [ fig : mse_mixture_vs_t ] , we display the mse of the aml estimator versus @xmath46 and versus @xmath70 for different values of @xmath253 . clearly , the rate of convergence of the estimator is affected by a small amount of gaussian noise , even when @xmath70 is small . this indicates that , if noise is not purely @xmath0-distributed with small @xmath70 , we recover the usual behavior of the mse versus @xmath46 . in this paper we addressed the doa estimation problem in @xmath0-distributed noise using two data sets . the main result of the paper was to show that , when the shape parameter @xmath70 of the texture gamma distribution is below @xmath254 , the fim is unbounded . on the other hand , for @xmath110 , the fim is bounded and we derived an accurate closed - form approximation of the crb . the maximum likelihood estimator was derived as well as an approximation , which induces non significant losses compared to the exact mle . in the non regular case where @xmath244 , we derived lower and upper bounds on the mean - square error of the ( a)ml estimates and we showed that the rate of convergence of these ( a)ml estimates is about @xmath213 where @xmath46 is the number of snapshots . c. j. coleman . the directionality of atmospheric noise and its impact upon an hf receiving system . in _ proceedings 8th international conference hf radio systems techniques _ , pages 363366 , guildford , uk , 2000 . abramovich , g. san antonio , and g. j. frazer . over - the horizon radar signal - to -external noise ratio improvement in over - sampled uniform 2d antenna arrays : theoretical analysis of superdirective snr gains . in _ proceedings ieee radar conference _ , pages 15 , ottawa , canada , 29 april-3 may 2013 . abramovich and g. san antonio . over - the horizon radar potential signal parameter estimation accuracy in harsh sensing environment . in _ proceedings ieeee international conference acoustics speech signal processing _ , pages 801804 , florence , italy , 4 - 9 may 2014 . f. pascal , y. chitour , j .- ovarlez , p. forster , and p. larzabal . covariance structure maximum - likelihood estimates in compound gaussian noise : existence and algorithm analysis . , 56(1):3448 , january 2008 . m. n. el korso , a. renaux , and p. forster . under @xmath255-distributed observation with parameterized mean . in _ proceedings 8th ieee sensor array and multichannel signal processing workshop _ , pages 461464 , a corua , spain , 22 - 25 june 2014 . y. i. abramovich and n. k. spencer . diagonally loaded normalised sample matrix inversion ( lnsmi ) for outlier - resistant adaptive filtering . in _ proceedings icassp _ , pages 11051108 , honolulu , hi , april 2007 . y. i. abramovich and o. besson . regularized covariance matrix estimation in complex elliptically symmetric distributions using the expected likelihood approach - part 1 : the oversampled case . , 61(23):58075818 , december 2013 . o. besson and y. i. abramovich . regularized covariance matrix estimation in complex elliptically symmetric distributions using the expected likelihood approach - part 2 : the under - sampled case . , 61(23):58195829 , december 2013 .
we consider the problem of estimating the direction of arrival of a signal embedded in @xmath0-distributed noise , when secondary data which contains noise only are assumed to be available . based upon a recent formula of the fisher information matrix ( fim ) for complex elliptically distributed data , we provide a simple expression of the fim with the two data sets framework . in the specific case of @xmath0-distributed noise , we show that , under certain conditions , the fim for the deterministic part of the model can be unbounded , while the fim for the covariance part of the model is always bounded . in the general case of elliptical distributions , we provide a sufficient condition for unboundedness of the fim . accurate approximations of the fim for @xmath0-distributed noise are also derived when it is bounded . additionally , the maximum likelihood estimator of the signal doa and an approximated version are derived , assuming known covariance matrix : the latter is then estimated from secondary data using a conventional regularization technique . when the fim is unbounded , an analysis of the estimators reveals a rate of convergence much faster than the usual @xmath1 . simulations illustrate the different behaviors of the estimators , depending on the fim being bounded or not .
after observation of the smallest neutrino oscillation by daya bay @xcite for the first time at the c.l . of @xmath1 , and immediately further confirmed by reno @xcite , dozens of solar , atmospheric , reactor , and accelerator - based neutrino experiments have compellingly provided us with solid information on five physical parameters involving three active flavors of massive neutrinos @xcite . the three mixing angles quantify the difference between the flavor and mass eigenstates of the three generations of neutrinos and the amplitude of neutrino oscillations ; the two independent mass squared differences are directly connected to neutrino mass spectrum and crucial to the oscillation frequencies in neutrino experiments . although our knowledge on neutrinos in the standard model ( sm ) of particle physics has achieved great progress , many basic properties of these magic particles are unclear and to be explored , e.g. , their absolute masses , the mechanism for their tiny masses , and cp violation in the lepton sector . one puzzle on neutrinos is the long - standing lsnd anomaly @xcite , which points to the existence of one ev scale neutrino and can not be accommodated in the three - neutrino paradigm . the anomalous @xmath2 in lsnd motivated the miniboone experiment @xcite , whose latest low energy data support the lsnd result , although the neutrino and antineutrino mode of the miniboone experiment are incompatible to a large extent @xcite . furthermore , recalculation of the reactor @xmath3 flux indicates the anomalous @xmath3 disappearance @xcite , i.e. , the so - called `` reactor anomaly '' . the gallium radioactive source experiments gallex and sage @xcite have also reported hints of @xmath4 disappearance , corresponding to the ev scale neutrino oscillation . in addition , cosmological data on cosmological microwave background , large scale structure , and the hubble constant imply extra light degrees of freedom @xcite , and light extra neutrinos beyond the three in the sm are convincingly the most natural candidate . all these hints of extra neutrinos , however , are restricted by , e.g. , the icarus@xcite and @xmath5 scattering @xcite data . due to the stringent limit on the @xmath6 boson decay width @xcite , the extra light neutrinos can not directly participate in the weak interaction and is consequently sterile . the simplest extension of the lepton sector is the addition of one sterile flavor to the three active ones , i.e. , the resultant ( 3 + 1 ) neutrino model . in this paper we concentrate on the simplest extension , although introducing more sterile states may reconcile comparatively better the data sets on sterile neutrinos and obtain a more preferable goodness - of - fit @xcite . neutrino mass matrix @xmath0 with texture zeros , i.e. , vanishing elements , is an attractive framework to phenomenologically study neutrinos and the connection of neutrino physics to sm and new physics beyond sm . texture zero structures imply the underlying flavor symmetry or mechanism @xcite , e.g. , the @xmath7 symmetry or froggatt - nielsen mechanism @xcite . furthermore , with vanishing elements in @xmath0 , the physical parameters on neutrino mass , mixing and cp violation are always connected to each other . the framework of texture zeros have been intensively studied in the three - neutrino frame @xcite and it can likewise apply to the ( 3 + 1 ) model with an extra sterile neutrino @xcite . in the ( 3 + 1 ) model , if all the four neutrino states are of majorana type , the neutrino mass matrix @xmath0 has ten independent elements , @xmath8 compared to six in the three - neutrino scheme . ghosh _ et al . _ @xcite has recently investigated the ( 3 + 1 ) model with two zeros , analyzing systematically all the @xmath9 distinct textures . they found that the 15 textures with both the two zeros in the active sector ( the first three rows and columns ) of @xmath0 are compatible with current neutrino data , whereas the 30 with zero(s ) in the sterile sector ( the fourth row and column ) are excluded by experiments mainly due to the mass hierarchy between the active and sterile neutrinos . , as well as one matrix element in the active sector , vanishes , one resultant possibility is that the mixing angle @xmath10 is close to zero . as the short baseline experiments are not sensitive to the tauon flavor , the vanishing of @xmath10 is allowed @xcite and it is yet immature to assert that such textures are excluded . see the following analytical and numerical discussions for details . ] in the presence of cp violation phases , one zero in @xmath0 generally imposes two correlation conditions on the parameters on neutrino masses , mixing angles and cp violating phases , i.e. , one is related to the vanishing absolute value and another to the vanishing argument . in general , more zeros means more stringent constraints on neutrino parameters and consequently more accurate predictions . in this work we focuses on @xmath11 neutrino mass matrices with three vanishing elements , assuming all the four neutrino states are of majorana type . as a comparison , none of the 20 @xmath12 mass matrices with three zeros are compatible with current data @xcite . adding one sterile neutrino , the number of possible textures with three zeros sum up to @xmath13 and this work aims to investigate the feasibility of three zero textures in the ( 3 + 1 ) scheme and explore their phenomenological implications . some elements of the lepton mixing matrix are more lengthy in the ( 3 + 1 ) model than in the three neutrino scheme @xcite and resultantly most of the elements of @xmath0 become even more lengthy . however , due to the large mass hierarchy between the sterile and active neutrinos @xmath14 and the smallness of the angles governing reactor mixing @xcite and active - sterile mixing ( asm ) @xcite , the vanishing conditions @xmath15 ( @xmath16 ) can be unbelievably simplified and we arrive at clean expressions for the three asm angles . based on these expressions we predict the restricted ranges of the asm angles and other phenomenological observables , e.g. , the effective majorana electron neutrino mass @xmath17 in neutrinoless double beta decay ( @xmath18 ) . we also numerically analyse these textures and study their phenomenological implications . we apply experimental constraints on the five active neutrino oscillation parameters , making use of the @xmath19 ranges from ref . @xcite , and examine which textures can accommodate the ev scale sterile neutrino mass and @xmath200.1 asms implied by short baseline ( sbl ) experiments . the sterile neutrino fits @xcite are somehow different from each other due to variations of the experimental data and calculation procedure , and we assign the fit in ref . @xcite as the criterion on the three parameters : lsnd mass square difference @xmath21 , the asm angle @xmath22 relating @xmath4 and @xmath23 , and @xmath24 relating @xmath25 and @xmath23 . specifically , as cosmological data prefer lighter neutrino states , we use the @xmath19 parameter regions in the @xmath26 and @xmath27 planes with @xmath28 @xmath21 in fig . 1 of ref . @xcite , where the effective mixing angles are defined as @xmath29 with @xmath30 . sbl experiments are not sensitive to the neutrino of tauon flavor , then the constraint on the third asm angle @xmath10 is comparatively rather loose , @xmath31 @xcite . in our numerical calculation , all the randomly generated parameters constrained by experiments are evenly distributed in their @xmath19 ranges . other parameters , e.g. , the cp violating phases @xmath32 , have not been determined at the @xmath19 c.l . if they are randomly generated , we assume they distribute evenly in their logically allowed ranges , e.g. , @xmath33 $ ] . the paper is organized as follows : the 120 textures with three zeros in the ( 3 + 1 ) model are systematically classified and labeled in the appendix . the formulas and experimental constraints on the neutrino parameters are briefly exposited in sec . [ sec : notation ] . section [ sec : analytical ] is devoted to analytical analysis of the textures and sec . [ sec : calculation ] to the numerical discussions . we briefly comment on the textures without cp violation in sec . [ sec : cpconservation ] and end up with the conclusion in sec . [ sec : conclusion ] . before demonstration of the analytical and numerical calculations , it is necessary to clarify our notations of the neutrino parameters and briefly explain relevant formulas , which is the aim of this section . we also sketch the current experimental constraints on the neutrino parameters . assuming the charged lepton mass matrix is diagonal and positive - definite , the @xmath11 majorana neutrino mass matrix @xmath34 can be expressed as , in the flavor basis , @xmath35 where @xmath36 is the diagonal neutrino matrix in the mass basis , and @xmath37 is the @xmath11 pmns lepton mixing matrix . in general , @xmath37 can split into two parts , @xmath38 @xmath39 is the lepton mixing matrix for dirac neutrinos , containing the solar mixing angle @xmath40 , atmospheric mixing angle @xmath41 , reactor mixing angle @xmath42 , three asm angles @xmath22 , @xmath24 , and @xmath10 , and three dirac cp violating phases . if we parametrize @xmath39 explicitly as @xcite @xmath43 the three dirac phases are @xmath32 , @xmath44 , and @xmath45 , where @xmath46 is the rotation matrix in the _ ij _ flavor space , e.g. , @xmath47 with @xmath48 and @xmath49 . the matrix @xmath50 in eq . ( [ matrix : v ] ) is a diagonal majorana phase matrix written as @xmath51 the factors @xmath52 s are introduced as a consequence of the mixing matrix @xmath37 appearing twice in eq . ( [ mnu ] ) . noticing that @xmath53 is sandwiched by the diagonal phase matrix @xmath50 , we define [ lambda ] @xmath54 and the @xmath11 neutrino mass matrix is slightly simplified to @xmath55 when three independent entries of @xmath0 are set to zero , [ 3zero ] @xmath56 where @xmath57 , or , explicitly , [ lambdan ] @xmath58 as we have only three complex equations , less than the number of neutrino flavors in the ( 3 + 1 ) model , we divide @xmath59 by @xmath60 and obtain the three following quantities : @xmath61 [ lambdann ] with three zeros in the @xmath11 matrix @xmath0 , based on eqs . ( [ lambdan ] ) and ( [ lambdann ] ) , @xmath62 can be expressed as functions of the matrix elements of @xmath39 , @xmath63 [ lambdannn ] where @xmath64 in presence of the three dirac cp violating phases @xmath32 , @xmath44 and @xmath45 , the matrix @xmath39 is generally complex , then , from the definitions of @xmath62 , we can extract three independent neutrino mass ratios [ massratio ] @xmath65 and the three majorana cp violating phases [ majoranaphase ] @xmath66 given the three vanishing conditions on @xmath0 , i.e. , eq . ( [ 3zero ] ) , the neutrino mass ratios and majorana phases are related to the six lepton mixing angles and the three dirac cp violating phases . to further relate these neutrino parameters to neutrino oscillation experiments and utilize the experimental constraints on the solar neutrino mass square difference @xmath67 , the atmospheric neutrino mass square difference @xmath68 as well as @xmath21 , we define the two independent neutrino mass square difference ratios [ rnu ] @xmath69 for the convenience of the analytical and numerical calculations below , let us first clearly count the relevant neutrino parameters we are confront with . in presence of the extra sterile flavor , we have altogether 16 physical parameters * six mixing angles between different flavors : three denote the mixing between active flavors , @xmath70 , while the other three are relevant to the asms , @xmath71 . * six cp violating phases : three are of dirac type , @xmath72 , and the other three are of majorana type , @xmath73 , @xmath74 , and @xmath75 . * four neutrino masses : @xmath76 , @xmath77 , @xmath78 , and @xmath79 , with the first three predominantly of active flavors , and the last one predominantly of sterile . currently , the three active neutrino mixing angles and two mass squared differences @xmath67 and @xmath68 can be extracted from neutrino oscillation data . their corresponding central values and @xmath19 ranges are listed in table [ table : nu - fit ] . the dirac cp violating phase @xmath32 in the ( 3 + 1 ) model corresponds to the sole dirac phase @xmath80 in the three - neutrino scheme . although it can be loosely constrained in global neutrino fits , we treat it as a free parameter due to the low sensitivity of current data to cp violation @xcite . .experimental data on neutrino oscillation parameters @xcite . [ cols="^,^,^,^",options="header " , ] + in the texture @xmath81 ( nh , upper left panel ) , @xmath45 in @xmath82 ( nh , upper right panel ) , @xmath73 in @xmath83 ( ih , middle left panel ) and @xmath84 ( nh , middle right panel ) , @xmath74 in @xmath85 ( nh , lower left panel ) and @xmath75 in @xmath84 ( nh , lower right panel ) , as functions of the lightest neutrino mass @xmath86 . the blue and red points denote , respectively , the nh and ih spectrums.,title="fig:",width=230 ] in the texture @xmath81 ( nh , upper left panel ) , @xmath45 in @xmath82 ( nh , upper right panel ) , @xmath73 in @xmath83 ( ih , middle left panel ) and @xmath84 ( nh , middle right panel ) , @xmath74 in @xmath85 ( nh , lower left panel ) and @xmath75 in @xmath84 ( nh , lower right panel ) , as functions of the lightest neutrino mass @xmath86 . the blue and red points denote , respectively , the nh and ih spectrums.,title="fig:",width=230 ] + in the texture @xmath81 ( nh , upper left panel ) , @xmath45 in @xmath82 ( nh , upper right panel ) , @xmath73 in @xmath83 ( ih , middle left panel ) and @xmath84 ( nh , middle right panel ) , @xmath74 in @xmath85 ( nh , lower left panel ) and @xmath75 in @xmath84 ( nh , lower right panel ) , as functions of the lightest neutrino mass @xmath86 . the blue and red points denote , respectively , the nh and ih spectrums.,title="fig:",width=230 ] in the texture @xmath81 ( nh , upper left panel ) , @xmath45 in @xmath82 ( nh , upper right panel ) , @xmath73 in @xmath83 ( ih , middle left panel ) and @xmath84 ( nh , middle right panel ) , @xmath74 in @xmath85 ( nh , lower left panel ) and @xmath75 in @xmath84 ( nh , lower right panel ) , as functions of the lightest neutrino mass @xmath86 . the blue and red points denote , respectively , the nh and ih spectrums.,title="fig:",width=230 ] + in the texture @xmath81 ( nh , upper left panel ) , @xmath45 in @xmath82 ( nh , upper right panel ) , @xmath73 in @xmath83 ( ih , middle left panel ) and @xmath84 ( nh , middle right panel ) , @xmath74 in @xmath85 ( nh , lower left panel ) and @xmath75 in @xmath84 ( nh , lower right panel ) , as functions of the lightest neutrino mass @xmath86 . the blue and red points denote , respectively , the nh and ih spectrums.,title="fig:",width=230 ] in the texture @xmath81 ( nh , upper left panel ) , @xmath45 in @xmath82 ( nh , upper right panel ) , @xmath73 in @xmath83 ( ih , middle left panel ) and @xmath84 ( nh , middle right panel ) , @xmath74 in @xmath85 ( nh , lower left panel ) and @xmath75 in @xmath84 ( nh , lower right panel ) , as functions of the lightest neutrino mass @xmath86 . the blue and red points denote , respectively , the nh and ih spectrums.,title="fig:",width=230 ] * as the analytical analysis , the numerical calculation can also predict the physical nonoscillation observables : sum of neutrino masses @xmath87 , the effective neutrino mass @xmath88 in single beta decay and the neutrino mass @xmath17 in @xmath18 . as an explicit illustration , the three quantities for the texture @xmath89 are presented in fig . [ fig : non - oscillation ] . the three light active neutrinos are generally of order 0.01 ev and the sterile one of order 1 ev , thus the masses of the four states sum up to @xmath201 ev , as shown in the upper panel of fig . [ fig : non - oscillation ] , which is compatible with the bound on neutrino mass from cosmological observations @xcite . + numerical prediction of the effective neutrino mass @xmath88 in tritium beta decay for the texture @xmath89 is presented in the lower left panel of fig . [ fig : non - oscillation ] . considering the sbl constraints , the lightest neutrino mass @xmath86 can not too larger than 0.01 ev , or the asms would become very large . thus , assuming @xmath90 ev , the effective neutrino mass @xmath88 is around @xmath91 , consistent with the analytical prediction in sec . [ sec : analytical ] . a more phenomenologically meaningful output is that the upper bound on @xmath88 for the nh spectrum is below the katrin discover sensitivity of 0.35 ev @xcite , as indicated by the lo analytical expressions . if the electron neutrino mass @xmath88 is above 0.35 ev , katrin can in principle distinguish the two neutrino spectrums . however , the parameter space for the nh spectrum with @xmath90 ev and @xmath92 is rather small and it is challenging for kartrin to produce positive signal in this restricted region and finally pin down the neutrino spectrum . if the katrin experiment can not detect electron neutrino mass but otherwise set a upper limit of 0.2 ev , then we are not able to differentiate the two neutrino hierarchies . + for @xmath18 , as shown in the lower right panel of fig . [ fig : non - oscillation ] , with @xmath93 ev , the effective majorana neutrino mass @xmath17 for ih spectrum is large than @xmath20 40 mev , consistent with the analytical predictions . the @xmath17 for nh spectrum is clearly much lower , predictively around 10 mev . the manifestation of nonvanishing lower limit for nh , which seems not consistent with the lo prediction , is a result from the corrections of the nlo and higher order terms . as mentioned in sec . [ sec : analytical ] , its is promising for the next generation @xmath18 experiments to pin down the neutrino mass spectrum and test the type ii textures @xcite . + as functions of the lightest neutrino mass @xmath86 : sum of neutrino masses @xmath87 ( upper panel ) , effective neutrino mass @xmath88 ( lower left panel ) in beta decay and the effective majorana neutrino mass @xmath17 ( lower right panel ) in @xmath18 . the blue filled circle and red square denote , respectively , the nh and ih spectrums.,title="fig:",width=238 ] + as functions of the lightest neutrino mass @xmath86 : sum of neutrino masses @xmath87 ( upper panel ) , effective neutrino mass @xmath88 ( lower left panel ) in beta decay and the effective majorana neutrino mass @xmath17 ( lower right panel ) in @xmath18 . the blue filled circle and red square denote , respectively , the nh and ih spectrums.,title="fig:",width=238 ] as functions of the lightest neutrino mass @xmath86 : sum of neutrino masses @xmath87 ( upper panel ) , effective neutrino mass @xmath88 ( lower left panel ) in beta decay and the effective majorana neutrino mass @xmath17 ( lower right panel ) in @xmath18 . the blue filled circle and red square denote , respectively , the nh and ih spectrums.,title="fig:",width=238 ] the lo analytical analysis indicates that there exist approximate @xmath94 symmetry between some of the type ii textures . in the numerical calculation without any approximation , the approximate @xmath94 symmetry are reserved and the textures related by this flavor symmetry share the same phenomenological features , e.g. , the preference of the textures @xmath95 for the nh spectrum over ih . one antiexample is that the texture @xmath84 prefer the cp violating phases @xmath96 and @xmath97 but its partner @xmath89 does not , as presented in table [ table : phases ] . we speculate this explicit and significant flavor `` asymmetry '' is due to the nlo and higher order corrections beyond eq . ( [ eqn : ii ] ) . when all the cp violating phases vanish , i.e. , @xmath98 , without tuning of the phases it would be more challenging for the vanishing conditions ( [ 3zero ] ) to be satisfied . for instance , the cp conserving partners of the six vanishing conditions in the active sector of @xmath0 are , respectively , @xmath99 obviously , the first four equations can never be satisfied , as the left - hand sides of them are definitely positive . consequently , none of the 20 type ii textures without cp violation can even be compatible with the current data on active neutrino oscillations . we calculate numerically all these 20 cp conserving textures and predictively obtain the negative result . for the 100 type i cp conserving textures , due to the same reason above for the type ii textures , they are all expected to be excluded by the active neutrino data . we also numerically analyze these 100 cp conserving textures and again arrive at the same conclusion . the lsnd anomaly , miniboone experiment , the reactor anomaly , the gallium anomaly and cosmological observations all point to the existence of light ev scale sterile neutrino(s ) . texture zero is a powerful framework to investigate the phenomenologies of the neutrino sector of the sm . in this work we apply the framework of texture zeros to the smallest extension of the neutrino sector of sm , i.e. , the ( 3 + 1 ) model , and examine all the 120 textures of @xmath11 majorana neutrino mass matrix @xmath0 with three zeros . although all the @xmath12 neutrino matrices with three zeros have been excluded , we find that when the neutrino sector is extended to include one sterile flavor , the situation is completely different . we classify the 120 textures into two large classes , the 100 type i textures with zeros in the sterile sector ( the fourth row and column ) of @xmath0 and the 20 type ii textures without zero in the sterile sector . with the approximations of large mass hierarchy @xmath100 and smallness of mixings @xmath101 , all the analytical expressions can be greatly simplified . we analyze both analytically and numerically all the 120 textures and find that the 100 type i textures are all excluded by current neutrino data , mainly due to the large active - sterile mass splitting . the exception among the type i textures are the class @xmath102 textures with only @xmath103 in the sterile sector , but they are highly disfavored by experiments as a result of the extremely small available parameter space . the 20 type ii textures with all the three zeros in the active sector are phenomenologically much more favored . as for the textures within the three - neutrino scheme @xcite , some of the textures are related by the approximate @xmath94 symmetry @xcite and the symmetric partners always share the same phenomenological features . their remarkable features revealed by both the analytical and numerical analysis are collected as follows : 1 . to obtain vanishing elements in the active sector of @xmath0 and keep compatible with the @xmath104 active - sterile mixing , the active neutrino mass and lightest neutrino mass are required to be of order 0.01 ev . 2 . we can also state the same fact in another way : if the active neutrinos are of order 0.01 ev , we naturally arrive at the asms are of order 0.1 . that is to say , the 20 textures are overall compatible with current experimental data , which is verified by both analytical and numerical analysis . numerical calculations reveal further that among the 20 textures , only the texture @xmath105 is actually excluded and the other 19 are more or less compatible with data . both the two neutrino spectrums of nh and ih for the three light active flavors are allowed in the textures @xmath83 , @xmath106 , @xmath107 , @xmath108 , @xmath81 , whereas the seven textures @xmath95 , @xmath109 , @xmath110 , @xmath111 prefer much for the nh spectrum over ih . 4 . some of the cp violating phases in the 19 textures exhibit preference of particular values , e.g. , the dirac phase @xmath32 in the texture @xmath81 favors the values around @xmath112 or @xmath113 , and in @xmath83 the majorana phase @xmath114 . all the predictions on the phases are collected in table [ table : phases ] . the dirac - type phases , especially the phase @xmath32 , can be observed in the upcoming long baseline neutrino oscillation experiments , while the majorana phases are crucial to the @xmath18 experiments . all these relevant oscillation and nonoscillation experiments are promising to examine the variant phenomenological implications and test these textures in the ( 3 + 1 ) model . the 0.01 ev active neutrinos and ev scale sterile neutrino sum up to @xmath20ev , compatible with the cosmological bound on neutrino mass . no matter whether the katrin beta decay experiment can observe the electron neutrino mass or not , it is challenging for it to distinguish the two neutrino spectrums in the three - zero textures . in contrast , it is promising for the next generation @xmath18 experiments with sensitivity of order 10 mev can clearly differentiate nh from ih and then test these textures . for example , if the effective majorana neutrino mass comes up to be 30 mev , the three active neutrino are probably of inverted hierarchy and then those textures preferable of nh would be almost excluded . test of the textures in this item is based on the effective neutrino mass , while the @xmath18 test in the previous item is related to the majorana phase detection in @xmath18 experiments . although both the tests are in principle feasible , the latter one is much more challenging than the former @xcite . ] in addition to the textures with the most general cp violation , we calculate also the textures without any cp violation and find that without the tuning of the phases in the correlation equations , none of the 120 textures are allowed by current neutrino data . in the end we stress that with more accurate neutrino data , especially the determination of the octant of @xmath41 @xcite , the textures with three zeros in the ( 3 + 1 ) model will be more severely constrained . it is likely that more textures will be excluded and we can obtain more accurate phenomenological predictions , e.g. , the effective neutrino mass in the neutrinoless double beta decay . the author would like to thank wei chao for the frequent discussions in the early stage of this work . the author also appreciates zhi - zhong xing , rabindra mohapatra , and guo - zhu ning for reading the manuscript and the valuable comments by xing and ning . this work is supported by national natural science foundation of china ( nsfc ) under grant no . the number of textures for @xmath11 @xmath0 with three zeros is rather large ; thus , it is beneficial for us to properly classify and label the 120 distinct textures . as the sterile neutrino is essentially different from the three active ones , the 120 textures can be roughly classified into two large categories : ( i ) 100 textures with zeros in the sterile sector of @xmath0 and ( ii ) 20 textures without vanishing elements in the sterile sector . as we will see in the analytical and numerical discussions , predictions of the structures in the two classes on the asm angles are completely different . for the convenience of discussions , the 100 textures of @xmath0 of class i can be further classified into six subgroups @xmath115 , according to the number of zeros in the sterile sector . the matrices given below are the sample matrices for the six subgroups : @xmath116 all the three texture zeros of the first two groups @xmath117 and @xmath118 are in the sterile sector . the three textures of type @xmath117 have the fourth diagonal element @xmath119 , whereas all the three asm angles of the texture of type @xmath118 are zeros and this texture is apparently excluded by the lsnd experiment . the two following groups @xmath120 and @xmath121 consist of the @xmath0 structures with one zero in the active sector and the other two in the sterile sector . @xmath119 holds true for the textures of type @xmath120 , while for the textures of type @xmath121 , the two texture zeros in the sterile sector are both nondiagonal elements . for the last two groups @xmath102 and @xmath122 , only one of the three texture zeros are from the sterile sector . for type @xmath102 @xmath119 , whereas for type @xmath122 one of the three nondiagonal elements in the fourth row or column vanishes . the 20 patterns of @xmath0 with all the three zeros in the active sector can also be further classified into subgroups . @xmath12 neutrino matrices with three texture zeros have been classified according to the positions of zeros , or more specifically according to the number of vanishing diagonal entries @xcite . the 20 @xmath11 matrices are analogically classified into six subgroups and they are explicitly labeled as follows : @xmath123 @xmath124 @xmath125 @xmath126 @xmath127 @xmath128 all the first three diagonal elements of the texture * a * vanish , whereas two of the first three diagonal elements of the textures of types * b * and * c * are zeros , only one for types * d * and * e * are vanishing and none of the three diagonal elements in the active part of * f * are zero . the three textures of type * c * are of rank three , which are distinguishable from the six of type * b * which are of rank four . similarly , the six structures of type * d * are of rank four , whereas the three of type * e * are of rank three . f. p. an _ et al . _ [ daya - bay collaboration ] , phys . lett . * 108 * , 171803 ( 2012 ) [ arxiv:1203.1669 [ hep - ex ] ] . j. k. ahn _ [ reno collaboration ] , phys . lett . * 108 * , 191802 ( 2012 ) [ arxiv:1204.0626 [ hep - ex ] ] . j. beringer _ [ particle data group collaboration ] , phys . d * 86 * , 010001 ( 2012 ) . c. athanassopoulos _ et al . _ [ lsnd collaboration ] , phys . lett . * 75 * , 2650 ( 1995 ) [ nucl - ex/9504002 ] ; phys . c * 54 * , 2685 ( 1996 ) [ nucl - ex/9605001 ] ; phys . lett . * 77 * , 3082 ( 1996 ) [ nucl - ex/9605003 ] ; a. aguilar - arevalo _ et al . _ [ lsnd collaboration ] , phys . d * 64 * , 112007 ( 2001 ) [ hep - ex/0104049 ] . a. a. aguilar - arevalo _ et al . _ [ miniboone collaboration ] , phys . rev . * 102 * , 101802 ( 2009 ) [ arxiv:0812.2243 [ hep - ex ] ] ; phys . lett . * 105 * , 181801 ( 2010 ) [ arxiv:1007.1150 [ hep - ex ] ] . a. a. aguilar - arevalo _ et al . _ [ miniboone collaboration ] , arxiv:1207.4809 [ hep - ex ] . t. .a . mueller _ et al . _ , phys . c * 83 * , 054615 ( 2011 ) [ arxiv:1101.2663 [ hep - ex ] ] ; g. mention _ et al . _ , phys . rev . d * 83 * , 073006 ( 2011 ) [ arxiv:1101.2755 [ hep - ex ] ] ; p. huber , phys . c * 84 * , 024617 ( 2011 ) [ arxiv:1106.0687 [ hep - ph ] ] . j. n. abdurashitov _ et al . _ , phys . c * 73 * , 045805 ( 2006 ) [ nucl - ex/0512041 ] . k. n. abazajian , m. a. acero , s. k. agarwalla , a. a. aguilar - arevalo , c. h. albright , s. antusch , c. a. arguelles and a. b. balantekin _ et al . _ , arxiv:1204.5379 [ hep - ph ] . mantonello , bbaibussinov , pbenetti , ecalligarich , ncanci , scentro , acesana and kcieslik _ et al . _ , arxiv:1209.0122 [ hep - ex ] . j. m. conrad and m. h. shaevitz , phys . d * 85 * , 013017 ( 2012 ) [ arxiv:1106.5552 [ hep - ex ] ] . s. schael _ et al . _ [ aleph and delphi and l3 and opal and sld and lep electroweak working group and sld electroweak group and sld heavy flavour group collaborations ] , phys . rept . * 427 * , 257 ( 2006 ) [ hep - ex/0509008 ] . g. karagiorgi , z. djurcic , j. m. conrad , m. h. shaevitz and m. sorel , phys . d * 80 * , 073001 ( 2009 ) [ erratum - ibid . d * 81 * , 039902 ( 2010 ) ] [ arxiv:0906.1997 [ hep - ph ] ] ; j. m. conrad , c. m. ignarra , g. karagiorgi , m. h. shaevitz and j. spitz , arxiv:1207.4765 [ hep - ex ] . m. maltoni and t. schwetz , phys . d * 76 * , 093005 ( 2007 ) [ arxiv:0705.0107 [ hep - ph ] ] ; j. kopp , m. maltoni and t. schwetz , phys . * 107 * , 091801 ( 2011 ) [ arxiv:1103.4570 [ hep - ph ] ] . w. grimus , a. s. joshipura , l. lavoura and m. tanimoto , eur . j. c * 36 * , 227 ( 2004 ) [ hep - ph/0405016 ] ; s. kaneko , h. sawanaka , t. shingai , m. tanimoto and k. yoshioka , hep - ph/0703250 [ hep - ph ] ; b. adhikary , a. ghosal and p. roy , jhep * 0910 * , 040 ( 2009 ) [ arxiv:0908.2686 [ hep - ph ] ] ; f. gonzalez canales and a. mondragon , j. phys . ser . * 287 * , 012015 ( 2011 ) [ arxiv:1101.3807 [ hep - ph ] ] ; s. dev , s. gupta and r. r. gautam , phys . b * 701 * , 605 ( 2011 ) [ arxiv:1106.3451 [ hep - ph ] ] . c. d. froggatt and h. b. nielsen , nucl . b * 147 * , 277 ( 1979 ) . z. -z . xing , hep - ph/0406049 . m. frigerio and a. y. .smirnov , phys . d * 67 * , 013007 ( 2003 ) [ hep - ph/0207366 ] ; a. dighe and n. sahu , arxiv:0812.0695 [ hep - ph ] ; z. -z . xing , phys . d * 68 * , 053002 ( 2003 ) [ hep - ph/0305195 ] ; z. -z . xing , phys . d * 69 * , 013006 ( 2004 ) [ hep - ph/0307007 ] ; a. merle and w. rodejohann , phys . d * 73 * , 073012 ( 2006 ) [ hep - ph/0603111 ] ; s. dev , s. verma and s. gupta , phys . lett . b * 687 * , 53 ( 2010 ) [ arxiv:0909.3182 [ hep - ph ] ] ; y. bentov and a. zee , phys . d * 84 * , 073012 ( 2011 ) [ arxiv:1103.2616 [ hep - ph ] ] ; e. i. lashin and n. chamoun , phys . d * 85 * , 113011 ( 2012 ) [ arxiv:1108.4010 [ hep - ph ] ] ; s. dev , s. gupta and r. r. gautam , phys . d * 82 * , 073015 ( 2010 ) [ arxiv:1009.5501 [ hep - ph ] ] . z. -z . xing , phys . b * 530 * , 159 ( 2002 ) [ hep - ph/0201151 ] ; z. -z . xing , phys . b * 539 * , 85 ( 2002 ) [ hep - ph/0205032 ] ; b. r. desai , d. p. roy and a. r. vaucher , mod . a * 18 * , 1355 ( 2003 ) [ hep - ph/0209035 ] ; w. -l . guo and z. -z . xing , phys . d * 67 * , 053002 ( 2003 ) [ hep - ph/0212142 ] ; r. mohanta , g. kranti and a. k. giri , hep - ph/0608292 ; s. dev , s. kumar , s. verma and s. gupta , nucl . b * 784 * , 103 ( 2007 ) [ hep - ph/0611313 ] ; s. dev , s. kumar , s. verma and s. gupta , phys . b * 656 * , 79 ( 2007 ) [ arxiv:0708.3321 [ hep - ph ] ] ; s. dev , s. kumar and s. verma , phys . d * 79 * , 033011 ( 2009 ) [ arxiv:0901.2819 [ hep - ph ] ] ; w. grimus and p. o. ludl , phys . b * 700 * , 356 ( 2011 ) [ arxiv:1104.4340 [ hep - ph ] ] ; p. o. ludl , s. morisi and e. peinado , nucl . b * 857 * , 411 ( 2012 ) [ arxiv:1109.3393 [ hep - ph ] ] ; d. meloni and g. blankenburg , arxiv:1204.2706 [ hep - ph ] ; w. grimus and p. o. ludl , arxiv:1208.4515 [ hep - ph ] ; p. h. frampton , m. c. oh and t. yoshikawa , phys . d * 66 * , 033007 ( 2002 ) [ hep - ph/0204273 ] ; g. bhattacharyya , a. raychaudhuri and a. sil , phys . d * 67 * , 073004 ( 2003 ) [ hep - ph/0211074 ] ; a. kageyama , s. kaneko , n. shimoyama and m. tanimoto , phys . b * 538 * , 96 ( 2002 ) [ hep - ph/0204291 ] ; m. honda , s. kaneko and m. tanimoto , jhep * 0309 * , 028 ( 2003 ) [ hep - ph/0303227 ] ; s. kumar , phys . d * 84 * , 077301 ( 2011 ) [ arxiv:1108.2137 [ hep - ph ] ] ; s. goswami and a. watanabe , phys . 033004 ( 2009 ) [ arxiv:0807.3438 [ hep - ph ] ] . z. -z . xing and s. zhou , phys . b * 593 * , 156 ( 2004 ) [ hep - ph/0403261 ] ; s. zhou and z. -z . xing , eur . j. c * 38 * , 495 ( 2005 ) [ hep - ph/0404188 ] . c. hagedorn and w. rodejohann , jhep * 0507 * , 034 ( 2005 ) [ hep - ph/0503143 ] ; w. -l . guo , z. -z . xing and s. zhou , int . j. mod . e * 16 * , 1 ( 2007 ) [ hep - ph/0612033 ] ; s. dev , s. kumar , s. verma and s. gupta , phys . d * 76 * , 013002 ( 2007 ) [ hep - ph/0612102 ] ; s. choubey , w. rodejohann and p. roy , nucl . b * 808 * , 272 ( 2009 ) [ erratum - ibid . * 818 * , 136 ( 2009 ) ] [ arxiv:0807.4289 [ hep - ph ] ] ; g. ahuja , m. gupta , m. randhawa and r. verma , phys . d * 79 * , 093006 ( 2009 ) [ arxiv:0904.4534 [ hep - ph ] ] ; s. goswami , s. khan and w. rodejohann , phys . b * 680 * , 255 ( 2009 ) [ arxiv:0905.2739 [ hep - ph ] ] ; x. -w . liu and s. zhou , arxiv:1211.0472 [ hep - ph ] . e. i. lashin and n. chamoun , phys . d * 80 * , 093004 ( 2009 ) [ arxiv:0909.2669 [ hep - ph ] ] ; h. a. alhendi , e. i. lashin and a. a. mudlej , phys . rev . d * 77 * , 013009 ( 2008 ) [ arxiv:0708.2007 [ hep - ph ] ] ; e. i. lashin and n. chamoun , phys . d * 78 * , 073002 ( 2008 ) [ arxiv:0708.2423 [ hep - ph ] ] . h. fritzsch , z. -z . xing and s. zhou , jhep * 1109 * , 083 ( 2011 ) [ arxiv:1108.4534 [ hep - ph ] ] ; y. shimizu , r. takahashi and m. tanimoto , arxiv:1212.5913 [ hep - ph ] . m. ghosh , s. goswami and s. gupta , arxiv:1211.0118 [ hep - ph ] . t. schwetz , talk at the conference sterile neutrinos at crossroads , virginia tech , usa , 2011 . z. -z . xing , int . j. mod . a * 19 * , 1 ( 2004 ) [ hep - ph/0307359 ] . g. karagiorgi , a. aguilar - arevalo , j. m. conrad , m. h. shaevitz , k. whisnant , m. sorel and v. barger , phys . d * 75 * , 013011 ( 2007 ) [ erratum - ibid . d * 80 * , 099902 ( 2009 ) ] [ hep - ph/0609177 ] . m. archidiacono , n. fornengo , c. giunti and a. melchiorri , phys . d * 86 * , 065028 ( 2012 ) [ arxiv:1207.6515 [ astro-ph.co ] ] . c. giunti and m. laveder , phys . d * 84 * , 073008 ( 2011 ) [ arxiv:1107.1452 [ hep - ph ] ] ; phys . d * 84 * , 093006 ( 2011 ) [ arxiv:1109.4033 [ hep - ph ] ] ; phys . b * 706 * , 200 ( 2011 ) [ arxiv:1111.1069 [ hep - ph ] ] ; c. giunti , m. laveder , y. f. li , q. y. liu and h. w. long , phys . rev . d * 86 * , 113014 ( 2012 ) [ arxiv:1210.5715 [ hep - ph ] ] . d. v. forero , m. tortola and j. w. f. valle , phys . d * 86 * , 073012 ( 2012 ) [ arxiv:1205.4018 [ hep - ph ] ] . j. barry , w. rodejohann and h. zhang , jhep * 1107 * , 091 ( 2011 ) [ arxiv:1105.3911 [ hep - ph ] ] . g. l. fogli , e. lisi , a. marrone , d. montanino , a. palazzo and a. m. rotunno , phys . d * 86 * , 013012 ( 2012 ) [ arxiv:1205.5254 [ hep - ph ] ] . m. auger _ [ exo collaboration ] , phys . lett . * 109 * , 032505 ( 2012 ) [ arxiv:1205.5608 [ hep - ex ] ] ; a. gando _ et al . _ [ kamland - zen collaboration ] , phys . lett . * 110 * , 062502 ( 2013 ) [ arxiv:1211.3863 [ hep - ex ] ] . j. fan and p. langacker , jhep * 1204 * , 083 ( 2012 ) [ arxiv:1201.6662 [ hep - ph ] ] . w. rodejohann , j. phys . g * 39 * , 124008 ( 2012 ) [ arxiv:1206.2560 [ hep - ph ] ] . f. m. fraenkle , arxiv:1110.0087 [ physics.ins-det ] . v. barger , r. gandhi , p. ghoshal , s. goswami , d. marfatia , s. prakash , s. k. raut and s u. sankar , phys . rev . lett . * 109 * , 091801 ( 2012 ) [ arxiv:1203.6012 [ hep - ph ] ] ; s. k. agarwalla , s. prakash and s. u. sankar , arxiv:1301.2574 [ hep - ph ] .
as a consequence of the lsnd anomaly and other hints of an ev scale sterile neutrino from particle physics and cosmology , the neutrino sector of the standard model of particle physics has to be extended and the smallest extension is the ( 3 + 1 ) model , i.e. , three active neutrinos plus one sterile one . in this work we study the neutrino mass matrix @xmath0 with three texture zeros in the ( 3 + 1 ) model , assuming all the neutrino states are of majorana type . with the mass hierarchy between the active and sterile neutrinos as well as the smallness of the reactor neutrino mixing and active - sterile mixing , the analytical expressions can be greatly simplified . we systematically examine all the 120 texture zeros , via both analytical and numerical analysis , and find that the 100 textures with zeros in the fourth column and row of @xmath0 can hardly be compatible with experiments ; whereas , 19 of the other 20 are feasible . we can foresee that the active neutrino masses are required to be around 0.01 ev and the next generation neutrinoless double beta decay experiments are promising to test these textures . many other phenomenologically meaningful predictions are also obtained . all the cp conserving ( 3 + 1 ) textures with three zeros are excluded by experiments .
a modern version of the constituent quark model with many attractive features is one of remarkable achievements of the schwinger - dyson ( sd ) approach ( reviewed in refs . @xcite ) to the physics of quarks and hadrons . it is well - known that most quark antiquark bound state approaches have grave problems when they are faced with the electromagnetic processes dominated by abelian axial anomaly . ( see ref . @xcite for a brief review and more references . ) however , these problems are resolved in the sd approach thanks to its good chiral properties . in the chiral limit , the axial - anomaly result for @xmath1 transition amplitude @xmath2 is reproduced analytically and exactly in this approach @xcite . the amplitude in eq . ( [ anomampl ] ) is the limit when both photons are real ( @xmath3 ) , of the general amplitude @xmath4 . at first , @xmath5 processes and other applications were studied in the approach with physically motivated _ ans " atze _ ( _ e.g. _ @xcite ) , instead of the solutions of sd equations ( sde ) with specified dynamics . however , the exact evaluation of abelian axial anomaly in the closed form ( [ anomampl ] ) occurs also in the variant where one actually solves sde for the quark propagators and , in a consistent approximation and with the same model interaction , bethe - salpeter ( bs ) equations ( bse ) for quark - antiquark bound states . this is the consistently coupled sd - bs approach , of which @xcite and references therein provide an elaborated example . since the anomaly is independent of hadronic structure , it is essential to note that the relation ( [ anomampl ] ) is successfully reproduced in the sd approach irrespective of what concrete _ ansatz _ for the quark propagator @xmath6^{-1}~ , \label{quark_propagator}\ ] ] is used , or what concrete consistently coupled sd - bs solutions are employed for the dressed quark propagator ( [ quark_propagator ] ) and the pion @xmath7 bound - state bs vertex @xmath8 . the successful treatment of the abelian axial anomaly is possible in the sd approach because this approach incorporates the dynamical chiral symmetry breaking ( d@xmath9sb ) into the bound states consistently , so that the pion , although constructed as a @xmath7 bound state , appears as a goldstone boson in the chiral limit . with that , the bound - state descriptions of mesons are reconciled with chiral requirements stemming both from qcd as underlying fundamental theory , and from phenomenology . the scenario is essentially of the nambu jona - lasinio ( njl ) type , but _ without _ its low cutoff . consequently , the quark - antiquark bound state mesons can finally provide adequate descriptions of the processes where the axial anomaly is important ( _ e.g. _ , see @xcite ) , such as the two - photon processes of light pseudoscalars , of which the @xmath10 decay is the cleanest example of an anomalous electromagnetic decay . having gotten in hands the fully satisfactory result ( [ anomampl ] ) for the @xmath10 , the next thing to explore was its one - off - shell - photon extension @xmath11 in the sd approach . @xcite studied it using _ ans " atze _ for dressed quark propagators ( [ quark_propagator ] ) , and then ref . @xcite did it in the consistently coupled sd - bs approach , utilizing the solutions of ref . however , both papers called for additional studies in two respects . first , both had employed the soft and chiral limit for the pion ; _ i.e. _ , the bs vertex was approximated by its leading , @xmath12 piece : @xmath13 ( @xmath14 is the third gell - mann matrix of flavor @xmath15 . ) in addition , refs . @xcite could compare their results just with then only available data by cello @xcite , at @xmath16{$<$ } } \put(.7,-0.3){\makebox(0.0,1.)[b]{$\sim$ } } \end{picture}2}.5 $ ] gev@xmath0 . lately , however , the interest in the form factor @xmath17 for the transition @xmath18 ( where @xmath19 is the momentum - squared of the spacelike off - shell photon @xmath20 ) , has again been growing for both experimental ( the new cleo data @xcite and plans for new tjnaf measurements @xcite ) and theoretical reasons _ e.g. _ , see @xcite . at large values of @xmath21 , the @xmath22 transition form factor should be adequately given by perturbative qcd ( pqcd ) - _ e.g. _ see @xcite . nevertheless , it is still not quite certain which @xmath21 is sufficiently large . for example , according to refs . @xcite , just pqcd may still be not quite sufficient even at the highest of the presently accessible momenta , @xmath16{$<$ } } \put(.7,-0.3){\makebox(0.0,1.)[b]{$\sim$ } } \end{picture}1}0 $ ] gev@xmath0 . the pqcd approaches start having problems as @xmath21 decreases more and more into the nonperturbative domain . also , see sec . iv of the recent ref . @xcite for clarifications how such approaches @xcite fail to reproduce the anomaly - induced value ( [ anomampl ] ) at @xmath23 . on the other hand , the treatment of the axial anomaly is a strong point of the sd approach , as pointed out above . what must then be explored in the coupled sd - bs approach is if the large @xmath21 behavior is satisfactory . in particular , the comparison with the new cleo data @xcite at @xmath21 up to 8 gev@xmath0 must be made . whether the large @xmath21 behavior is satisfactorily close to the data and to the predictions of pqcd is a tricky question for a constituent quark model . namely , the calculation of the transition form factor carried out in the simple constituent quark model ( with the constant light quark mass parameter @xmath24 ) , leads to @xmath25 as @xmath26 , which overshoots both the cleo data and pqcd predictions considerably @xcite . this is because of the additional @xmath27-dependence on top of the large @xmath21 behavior @xmath28 favored by pqcd and other qcd - based theoretical predictions such as @xcite and references therein . publication of the cleo data @xcite made it clear that the large-@xmath21 behavior ( [ largeq2 ] ) is also favored experimentally @xcite . fortunately , the constituent quark model provided by the sd approach , in which @xmath29 plays the role of the dynamically generated @xmath30-dependent mass , has turned out not to suffer from that shortcoming . namely , we have shown in ref . @xcite that the above asymptotic behavior ( [ largeq2 ] ) is in this approach obtained in the model - independent way . in this paper , we provide yet another way to obtain the asymptotic @xmath31 behavior found in ref . predictions for @xmath17 for finite values of @xmath21 can be obtained by picking a definite model for interactions between quarks and obtaining corresponding solutions of the sde and bse . in this paper , just as we did in ref . @xcite , we use the model of ref . however , here we give the results over much wider range of @xmath21 - see fig . 1 and the discussion thereof . comparison with the data reveals in what way more precise measurements at intermediate momenta , feasible at tjnaf , can both give insights in the hadronic structure and provide guidance how to improve presently existing models in the sd - bs approach . in addition , the coupled sd - bs variant of the sd approach ( such as that in ref . @xcite ) can readily be extended and applied beyond the soft and chiral limits [ given by eq . ( [ softlimbs ] ) ] for the bound - state vertex - _ e.g. _ see refs . we did this when calculating @xmath17 in ref . @xcite , but here we locate and point out the main reason for the difference between the full calculation of the transition form factor and the one using the the soft and chiral limit approximation ( [ softlimbs ] ) for the pion bs vertex . in the coupled sd - bs approach , the bse for the pion bound - state @xmath7 vertex @xmath33 employs the dressed quark propagator ( [ quark_propagator ] ) obtained by solving its sde . solving the sde and bse in a consistent approximation is crucial ( _ e.g. _ , see refs . @xcite ) for obtaining @xmath7 bound states which are , in the case of light pseudoscalar mesons , simultaneously also the ( pseudo-)goldstone bosons of d@xmath9sb . following jain and munczek @xcite , we adopt the ladder - type approximation sometimes called the _ improved _ @xcite or _ generalized _ @xcite ladder approximation ( employing bare quark gluon quark vertices but dressed propagators ) . for the gluon propagator we use an effective , ( _ partially _ ) modeled one in landau - gauge @xcite , given by @xmath34 ( this ansatz is often called the abelian approximation " @xcite . ) the effective propagator function @xmath35 is the sum of the perturbative contribution @xmath36 and the nonperturbative contribution @xmath37 : @xmath38 the perturbative part @xmath39 is well - known from perturbative qcd , so it is _ not _ modeled @xcite . as in refs . @xcite , we follow refs . @xcite and employ the two loop asymptotic expression for @xmath40 . for the modeled , ir part of the gluon propagator , we adopt from ref . @xcite @xmath41 , with their parameters @xmath42 and @xmath43 . for details of how we solve sde and bse , we refer to refs . @xcite . to high accuracy , we reproduce jain and munczek s @xcite solutions of sde for the dressed propagators @xmath44 , i.e. , the functions @xmath45 and @xmath46 , as well as the bs solutions for the four functions comprising the pion bound - state vertex @xmath8 . the @xmath47 transition tensor ( @xmath48 ) and scalar ( @xmath49 ) amplitudes , related by @xmath50 are found in the present paper as in refs . that is , the pseudoscalar - vector - vector ( pvv ) triangle graph is calculated by employing the framework advocated by ( for example ) @xcite in the context of electromagnetic interactions of bs bound states , and often called the generalized impulse approximation ( gia ) - _ e.g. _ , by @xcite . to evaluate the triangle graph , we therefore use the _ dressed _ quark propagator @xmath44 and the pseudoscalar ( @xmath51 ) meson bs bound state vertex @xmath52 . ( see fig . 1 in ref . @xcite , for example . ) equivalently , we can work with the bs - amplitude @xmath53 . this is in fact what we do , following ref . @xcite . the third ingredient crucial for gia s ability to reproduce the correct abelian anomaly result ( [ anomampl ] ) , is employing in the triangle graph appropriately dressed _ vector vertices @xmath54 , which satisfy the vector ward takahashi identity ( wti ) @xmath55 namely , assuming that photons couple to quarks through the bare vertex @xmath56 would be inconsistent with our quark propagator @xmath44 , which , dynamically dressed through its sd - equation , contains the momentum - dependent functions @xmath45 and @xmath46 . the bare vertex @xmath56 obviously violates the vector wti , implying the non - conservation of the electromagnetic vector current and charge . solving the pertinent sd equation for the dressed quark photon quark ( @xmath57 ) vertex @xmath58 is still an unsolved problem , and using the realistic anstze for @xmath58 is presently the only practical way to satisfy the wti . following refs . @xcite ( for example ) , we can choose the ball chiu ( bc ) @xcite vertex ; _ i.e. _ , @xmath59 , @xmath60 where @xmath61 $ ] , for @xmath62 or @xmath63 . this particular solution of the vector wti reduces to the bare vertex in the free - field limit as must be in perturbation theory , has the same transformation properties under lorentz transformations and charge conjugation as the bare vertex , and has no kinematic singularities . it does _ not _ introduce any new parameters as it is completely determined by the dressed quark propagator @xmath44 . another wti - preserving choice for @xmath58 can be a vertex of the curtis pennington ( cp ) type , _ i.e. _ , @xmath64 where the the transverse addition @xmath65 is of the type @xcite @xmath66 two especially suitable _ ans " atze _ for the dynamical function @xmath67 , ensuring multiplicative renormalizability of fermion sde beyond the ladder approximation in qed@xmath68 , are @xmath69 ^ 2 \right\}~. \label{defdfunct}\ ] ] the original cp _ ansatz _ @xmath70 employed @xmath71 @xcite . we will use it in analytic calculations of @xmath72 , which are possible for @xmath23 and @xmath73 . however , in the numerical calculations , which are necessary for finite values of @xmath74 , we prefer to restrict ourselves ( besides the minimal bc vertex ) to the _ modified _ cp ( mcp ) vertex , @xmath75 , resulting from the choice @xmath76 in eq . ( [ defdelta ] ) . namely , as pointed out in ref . @xcite , the numerical calculation of @xmath72 employing @xmath75 vertices is free from certain numerical difficulties arising , in this application , from the cp denominator function @xmath77 . in contrast to the bc one , the mcp vertex is also consistent with multiplicative renormalizability , like the original cp vertex . in the present context , the important _ qualitative _ difference between the bc - vertex on one side , and the cp vertex as well as the modified , mcp vertex on the other side , will be that @xmath78 when _ both _ @xmath79 , whereas @xmath80 and @xmath81 as soon as _ one _ of the squared momenta tends to infinity . this turned out @xcite to lead to the same coefficient of the @xmath31 behavior ( [ largeq2 ] ) for the cp vertices and mcp vertices , namely @xmath82 , but a larger one for the bc - vertices . we have checked that the @xmath83 case , _ i.e. _ , the @xmath10 amplitude ( [ anomampl ] ) , is reproduced analytically employing the cp vertices and mcp vertices in the same way as when the bc vertices were employed in earlier applications , _ e.g. _ @xcite . in the case of @xmath84 , gia yields ( _ e.g. _ , see eq . ( 24 ) in ref . @xcite ) the amplitude @xmath85 : @xmath86 @xmath87 here , @xmath9 is the bs amplitude of both @xmath88 and @xmath89 pseudoscalar bound states : @xmath90 thanks to the isospin symmetry assumed here . this symmetry likewise enables us to continue suppressing flavor labels also on the quark propagators @xmath91 and @xmath57 vertices @xmath58 . we follow the conventions of ref . @xcite , including those for the flavor factors and flavor matrices @xmath92 . then , @xmath93 , so that the prefactor @xmath94 in eq . ( [ tmunu(2 ) ] ) is just the flavor trace @xmath95 where @xmath96 is the quark charge matrix . already in ref . @xcite ( where only the bc vertex was employed ) , the transition form factor @xmath17 was numerically evaluated ( for @xmath97 gev@xmath0 only ) employing - for the pion only - the soft and chiral limit ( [ softlimbs ] ) . in the subsequent work @xcite , we went beyond this approximation , using our complete solution for the bs vertex @xmath33 , _ viz . _ the bs amplitude @xmath98 , given by the decomposition into 4 scalar functions multiplying independent spinor structures . we calculated the transition form factor not only for the bc vertex , but also for the mcp vertex . the approximation we kept in ref . @xcite ( and which we still keep in the present paper ) , is discarding the second and higher derivatives in the momentum expansions of the sd solutions @xmath45 and @xmath46 and bs solutions @xmath99 . in ref . @xcite , @xmath17 was evaluated up to @xmath100 gev@xmath0 , which is roughly the limit of the presently accessible transferred momenta . in this paper , we evaluate the transition form factor up to 36 gev@xmath0 . we present curves for @xmath101 evaluated in various ways first in fig . 1 , depicting their evolution all the way to 36 gev@xmath0 . this squeezes the presently existing experimental points into the first quarter of the scale ; however , the issue of comparing and improving the agreement with the data at the accessible momenta will be returned to later , and now we want first to illustrate how various theoretical transition form factors at finite @xmath21 connect to the asymptotic behavior . in fig . 1 , the solid line is our prediction for @xmath101 evaluated with the bc vertex . it reaches the asymptotic behavior faster than other curves . already after @xmath102 gev@xmath0 , the tiny changes can not be observed in the solid line , but only in the corresponding numerical data , as the asymptotic value is approached from below . for comparison , the dash - dotted line gives the results obtained ( also with the bc vertex ) in the approximation ( [ softlimbs ] ) of soft and chiral limit as explained above . at @xmath102 gev@xmath0 , where the @xmath101 from the full calculation already practically reaches the asymptotic limit , the difference between it and @xmath103 from the approximation ( [ softlimbs ] ) , is somewhat above 10% . this difference gets bigger as @xmath21 grows , and asymptotically it reaches some 22% of the limit value of @xmath101 . however , this still means that the effect of the soft limit approximation on the transition form factor is much less than on the charge form factor . namely , maris and roberts @xcite pointed out that omitting the pseudovector components of the pion lead even to the wrong asymptotic behavior , as @xmath104 instead of @xmath105 , in the case of the charge form factor of the pion . in the present case of the transition form factor , the correct @xmath105 leading behavior is nevertheless obtained not only for the results of the full calculation , but also in the soft limit approximation ( [ softlimbs ] ) . it is just that the coefficient of @xmath105 is underestimated with respect to the full calculation . the dashed line was obtained in the same way as the solid line , but employing the mcp vertex . even at @xmath106 gev@xmath0 , the asymptotic behavior is obviously not yet reached in the case of the mcp vertex . however , later we will comment on how ref . @xcite showed analytically that the behavior given by eq . ( [ largeq2 ] ) must be reached at some point , although at much higher values of @xmath21 when the mcp vertex is used . how do our transition form factors agree with other theoretical approaches ? the vector meson dominance ( vmd ) model and the qcd sum rule approach @xcite give the transition form factor which is some 10% below our bc " @xmath17 for the _ presently _ largest accessible values of @xmath21 , _ i.e. _ , around 8 gev@xmath0 . therefore , in that region , our bc - results are between the uppermost line in fig . 1 ( the dotted constant line at @xmath107 ) denoting the asymptotic pqcd @xcite version ( @xmath108 ) of eq . ( [ largeq2 ] ) , and the results of vmd ( _ e.g. _ , see @xcite ) , the recent pqcd calculation by ref . @xcite as well as the qcd sum rule results of radyushkin and ruskov @xcite . @xmath101 of vmd and pqcd rise with @xmath21 , albeit with different rates , while that from the qcd sum rules of ref . @xcite starts almost imperceptibly falling after @xmath109 gev@xmath0 . the large-@xmath21 leading power - law behavior ( [ largeq2 ] ) was first derived from the parton picture in the infinite momentum frame _ e.g. _ , see ref . @xcite . in this and other similar pqcd approaches , the precise value of the coefficient of the leading @xmath105 term depends on the pion distribution amplitude @xmath110 which should contain the necessary nonperturbative information about the probability that a partonic quark carries the fraction @xmath111 of the total longitudinal momentum . _ e.g. _ , the well - known example of a broad " distribution @xmath112 ( proposed by chernyak and zhitnitsky @xcite motivated by sum - rule considerations ) leads to @xmath113 , but it is too large in the light of the latest data @xcite . in contrast , the asymptotic distribution @xmath114 ( depicted by dotted line in fig . 2 and favored by lepage and brodsky @xcite ) yields @xmath115 , resulting in the constant dotted line crossing the upper error bar of the presently highest-@xmath21 data point in fig . 1 . in the _ strict _ @xmath116 limit , every distribution amplitude must evolve into the asymptotic one , @xmath117 , if the effects of the pqcd evolution are taken into account @xcite . however , even at @xmath21-values larger than the presently accessible ones , other effects may still be more important than the effects of the pqcd evolution . this is the reason why other approaches and other forms of @xmath110 should be considered even when they do not incorporate the pqcd evolution . this is in line with radyushkin and ruskov s ( @xcite and their references therein ) pointing out desirability of having direct calculations of @xmath17 without _ a priori _ assumptions about the pion distribution amplitude @xmath118 . one can then consider the opposite procedure from the one which is standard in pqcd : from such direct calculations of @xmath17 , one can draw conclusions about the distribution amplitude @xmath118 . the form @xmath119 ^ 2 } \ , x^\zeta \ , ( 1-x)^\zeta \ , , \quad \zeta > 0 \ , , \label{genform}\ ] ] is suitable for representing various distribution amplitudes because it is relatively general @xcite : @xmath120 yields the distributions that are peaked " or narrowed " with respect to the asymptotic one ( @xmath121 ) , whereas @xmath122 gives the broadened " distributions which , however , now seems to be ruled out for the same reason as @xmath123 quoted above , since @xmath124 is ruled out empirically by cleo @xcite . namely , it is easy to see that eq . ( [ genform ] ) implies @xmath125 . in fig . 2 , we plot the distribution amplitudes of the form ( [ genform ] ) for the three cases that are the most interesting for the present discussion . as seen below , these cases correspond to the @xmath126-values equal to 1 ( dotted line ) , 1.5 ( solid line ) , and 2.5 ( dash - dotted line ) . at the highest presently accessible momenta , the asymptotic prediction ( @xmath108 ) is lowered by some 20% by the lowest order qcd radiative corrections @xcite , amounting to @xmath127 , which fits the cleo data well . of course , these qcd corrections mean that @xmath101 is not strictly constant , but according to ref . @xcite it rises towards @xmath107 so slowly that we can take it constant in practice . ( 19 ) and ( 18 ) from ref . @xcite lead to the line of empty squares in fig . 1 , that is , to @xmath101 which grows just 4% from @xmath128 gev@xmath0 to @xmath129 gev@xmath0 when the number of relevant flavors is set to @xmath130 . the full calculation @xcite , which employs the momentum - dependent @xmath131 @xcite , seems to give even slower variation of @xmath132 see their fig . 2 . ] similarly slow variation results from the sum - rule approach of radyushkin and ruskov @xcite , which yields the transition form factor which is , on the interval from 3 to 8 gev@xmath0 , quite close to those of the pqcd approach of brodsky _ et al . the sum - rule @xcite @xmath101 starts actually _ falling _ after @xmath133 gev@xmath0 , but so slowly that eq . ( [ largeq2 ] ) with the constant @xmath127 represents it accurately at the presently accessible values of @xmath21 . this , just like roughly the same @xmath134 associated to ref . @xcite , corresponds to a rather narrow distribution ( [ genform ] ) with @xmath135 , depicted by the dash - dotted line in fig . the leading large-@xmath21 behavior as in eq . ( [ largeq2 ] ) , was obtained also by manohar @xcite using the operator product expansion ( ope ) . according to his ope calculation , the coefficient in eq . ( [ largeq2 ] ) giving the leading term is @xmath82 , which is below our large-@xmath21 @xmath17 by @xmath136 when we use the bc vertex , but exactly coincides with the @xmath26 limit obtained when we use the cp or the mcp vertex @xcite and comments below . the coefficient @xmath82 is the lowest one still consistent with the form ( [ genform ] ) because it corresponds to @xmath137 . the pion distribution amplitude ( [ genform ] ) then becomes infinitely peaked delta function : @xmath138 . for @xmath139 gev@xmath0 , our bc " @xmath17 also behaves in excellent approximation as ( [ largeq2 ] ) , with @xmath140 . this would in the pqcd factorization approach correspond to @xmath110 ( [ genform ] ) with @xmath141 . on the other hand , our mcp " @xmath17 falls off faster than @xmath105 for even the largest of the @xmath21 values depicted in fig . however , it does not fall off much faster , as our mcp " @xmath101 at @xmath142 gev@xmath0 is only 6% , and at the huge @xmath129 gev@xmath0 is only 10% smaller than at @xmath128 gev@xmath0 ( roughly the highest presently accessible @xmath21 ) . moreover , ref . @xcite showed analytically that , generally , the @xmath105-behavior of eq . ( [ largeq2 ] ) is at some point reached in our approach , although fig . 1 shows that for the mcp @xmath57 vertices this can happen only at significantly higher @xmath21 than it happens for the bc @xmath57 vertices . the most important novelty in ref . @xcite is its result on the large-@xmath21 asymptotic behavior . namely , it implies that the modern version of the constituent quark model which is given by the coupled sd - bs approach , provides from @xmath23 to @xmath73 the description for @xmath143 which is independently of model details consistent not only with the abelian axial anomaly but also with the qcd predictions ( [ largeq2 ] ) for the leading large-@xmath21 behavior . in the sd - bs approach , @xmath144 is a quantity calculable through the straightforward application of the mandelstam - formalism expression @xmath145 \ , . \label{fpi}\ ] ] the model dependence is present in the ( successfully reproduced @xcite ) _ value _ of @xmath144 , but the derivation @xcite of the asymptotic _ forms _ ( [ largeq2 ] ) and ( [ ampwtilde ] ) below , is model - independent . they do not depend on what are the bound - state solutions , just like the chiral - limit axial - anomaly amplitude ( [ anomampl ] ) does nt . of special importance is also that the derivation in ref . @xcite of the asymptotic large-@xmath21 forms ( [ largeq2 ] ) and ( [ ampwtilde ] ) below , applies to both minkowski and euclidean space . our result @xcite on the asymptotic behavior of @xmath17 subsequently received further support from roberts @xcite , who generalized its derivation by taking into account renormalization explicitly . his derivation shows that the asymptotics of ref . @xcite obtained using the cp or mcp vertex , namely @xmath82 , must be precisely reproduced for any @xmath57 vertex which is ( like the cp or mcp ones ) consistent with multiplicative renormalizability . it is interesting that the asymptotic behavior for large negative @xmath146 predicted @xcite for the bare @xmath57 vertex and the dressed cp and mcp ones , is in exact agreement with the leading term predicted by manohar using ope @xcite . in this paper , we shed some light on this connection by providing the following alternative derivation of the asymptotic behavior : consider @xmath147 , the @xmath148-product of two quark vector currents @xmath149 , along the lines of ref . @xcite ( ch . 18 on ope ) , but evaluated between the pion state and the vacuum . we substitute for the quark propagator its leading light cone [ @xmath150 behavior @xmath151 ^ 2 } \ , , \label{lightcones}\ ] ] and utilize @xmath152 similarly as in the derivation of the large @xmath21-behavior in ref . the term which contains the symmetric tensor @xmath153 , is readily seen not to contribute to the @xmath154 vertex . on the other hand , in the term containing the antisymmetric levi - civita tensor @xmath155 and @xmath156 , one finds in the lowest order of the @xmath157-expansion the pion - to - vacuum matrix element of the axial current , @xmath158 defining the pion decay constant @xmath144 . the coordinate - space integration is reduced to @xmath159 putting all this together reveals that the lowest - order result in the @xmath157-expansion is just eq . ( [ largeq2 ] ) with @xmath134 being precisely 4/3 . we thus in another way recover the result of manohar @xcite , of ref . @xcite for the case when the bare ( @xmath160 ) or the dressed cp or mcp vertices are used , and of subsequent generalization @xcite thereof for any @xmath57 vertices consistent with the multiplicative renormalization . however , this asymptotic behavior @xmath161 lies some 20% below the central values of the largest-@xmath21 cleo data , and is more than one standard deviation away from them . nevertheless , manohar @xcite pointed out that his ope approach also indicates the existence of potentially large corrections to his leading term . let us now return to the bc vertex , which has been the most commonly used dressed @xmath57 vertex in the phenomenological sd - bs calculations so far . in conjunction with our sd - bs solutions , the usage of the bc vertex raises the asymptotic coefficient @xmath134 of eq . ( [ largeq2 ] ) from 4/3 resulting from the usage of the bare , cp and mcp vertices , to @xmath162 . the reason for this enhancement is that the bc vertex does not reduce to the bare vertex @xmath56 if the large momenta flow through just one , and not both , of its fermion legs . ( also note that the bc vertex is not consistent with multiplicative renormalizability @xcite , so that the arguments of ref . @xcite are not applicable to it . ) it was shown in ref . @xcite that when the bc vertex is used in our approach , the predicted asymptotic behavior is @xmath163 where @xmath164 is given by the same expression ( [ fpi ] ) as @xmath144 , except that the integrand is modified by the factor @xmath165 ^ 2/4 $ ] . in the special case of the solutions we use , the slow and moderate variation of @xmath45 , as well as its shape ( see fig . 3 in ref . @xcite ) permits the approximate factorization @xmath166 ^ 2}{4}. \label{approxfactor}\ ] ] this last approximation ( a rather rough one ) would imply @xmath167 , _ i.e. _ , @xmath168 gev . the more accurate eq . ( [ ampwtilde ] ) gives @xmath169 gev . this is practically indistinguishable from the asymptotics indicated by our full numerical calculation of the transition form factor , which reached @xmath170 gev at the largest numerically studied @xmath21 . it is of special importance that analytically obtained large-@xmath21 asymptotics ( [ ampwtilde ] ) , the derivation of which is valid in both minkowski and euclidean space , makes this smooth and accurate contact between our large-@xmath21 numerical results calculated with the euclidean solutions of the chosen model @xcite . the excellent agreement between the analytical and numerical results confirms the accuracy of our numerical methods and procedures ( employing the euclidean bound - state solutions @xcite ) used in this and earlier papers @xcite . another case of this agreement between the numerical results and the analytical considerations is provided by @xmath101 found in the soft and chiral limit approximation ( [ softlimbs ] ) . therefore , let us for a moment consider again this approximation , although we surpassed it already in ref . ( as a bonus , we also gain understanding of some subtle aspects in its previous applications . ) in fig . 1 , the results of the numerical calculation in that approximation , and with the bc vertex ( [ bc - vertex ] ) , are depicted by the dash - dotted line , which practically reaches ( from above ) the asymptotic behavior , although not as fast as the results without that approximation , depicted by the solid line . we now note that in the approximation ( [ softlimbs ] ) , the asymptotic behavior ( [ ampwtilde ] ) should be realized with @xmath144 and @xmath164 calculated in the _ same _ approximation . we will denote them by @xmath171 and @xmath172 , where @xmath173 is short for generalized pagels - stokar " . namely , it turns out @xcite that the soft limit approximation ( [ softlimbs ] ) , when applied to @xmath144 , is the same as the widely used pagels - stokar approximation @xcite _ except _ that they work with the restriction @xmath174 unlike the present approach . since we find @xmath175 mev , it appears that our dash - dotted curve should approach @xmath176 gev . in fact , at @xmath129 gev@xmath0 , @xmath177 gev . to understand what forced the dash - dotted curve below , let us remember that the factor @xmath178 appears in eq . ( [ softlimbs ] ) because in the chiral limit it turns out that the pion decay constant @xmath144 is precisely equal to the the normalization of the bs vertex @xmath179 @xcite . in ref . @xcite the transition form factor was thus found using in eq . ( [ softlimbs ] ) our chiral - limit value of the pion decay constant , @xmath180 mev , and to make contact with that reference we have evaluated in this paper the dash - dotted curve in the same way . however , from what has just been explained , it follows that when we use not only the chiral , but also the soft limit ( [ softlimbs ] ) , the normalization of the bs vertex is given by the pion decay constant calculated in the same approximation , @xmath181 , which in our adopted model has the value of 80.8 mev . the transition form factor of ref . @xcite was thus suppressed by the factor @xmath182 ; for the same reason , the dash - dotted line in the present fig . 1 approaches from above @xmath183 gev as the limiting value , so everything tallies . in the case of the bc vertex , the soft " leg adjacent to the pion bs - amplitude always contributes @xmath45 , which enhances the asymptotic value of @xmath101 . this is not so for the cp or mcp vertices . since they tend to the bare vertex as soon as the high momentum flows through one of their legs , @xmath144 does _ not _ get replaced by @xmath164 , and the bare " result , eq . ( [ largeq2 ] ) with @xmath82 , continues to hold for @xmath26 when the cp or mcp vertices are used . of course , the usage of the mcp vertex instead of the bare @xmath56 makes a considerable difference for the finite @xmath21 . let us therefore now focus our attention to fig . 3 , which shows @xmath101 found in our approach for both the bc - case ( solid curve ) and the mcp - case ( dashed curve ) for the interval of the transferred momenta between the presently highest accessible @xmath21 and @xmath23 . at @xmath23 , both bc and mcp vertices give the same amplitude ( [ anomampl ] ) for @xmath10 in the chiral limit . with growing @xmath21 , the bc- and mcp - curves in fig . 3 must nevertheless ultimately separate because of the difference @xmath184 of their large-@xmath21 limiting values . at the highest presently accessible momenta , @xmath185 gev@xmath0 , they differ by 9% , but as we go down in @xmath21 , we see that already for @xmath16{$<$ } } \put(.7,-0.3){\makebox(0.0,1.)[b]{$\sim$ } } \end{picture}4}$ ] gev@xmath0 , they practically coincide . this insensitivity on the choice of the wti - preserving @xmath57-vertex is very indicative . namely , the true solution for the @xmath186 vertex , the one which we try to imitate by the bc and mcp vertex _ ans " atze _ , is for @xmath16{$<$ } } \put(.7,-0.3){\makebox(0.0,1.)[b]{$\sim$ } } \end{picture}4}$ ] gev@xmath0 unlikely to give @xmath101 significantly different from the bc and mcp vertices , since at @xmath23 the transition amplitude must ( in the chiral limit ) again be ( [ anomampl ] ) , and at larger @xmath21 the difference should not be larger than that resulting from the bc and mcp vertices , as they lead to so very different @xmath101 as @xmath26 . ( one should also keep in mind that the multiplication by @xmath21 serves only for exposing the asymptotics more clearly , and that the difference is even smaller for @xmath17 proper , which must in every case decrease as @xmath105 . ) the weak sensitivity ( at least for not too large @xmath21 ) on the wti - preserving _ ansatz _ for @xmath186 means that high - precision measurements of @xmath17 will test unambiguously ( and possibly give a hint on how to improve ) the sd and bs model solutions which have so far been successful in fitting the low - energy hadron properties such as the meson spectrum . for example , our model choice @xcite obviously ( for both bc and mcp vertices ) slightly overestimates @xmath17 in the region @xmath16{$<$ } } \put(.7,-0.3){\makebox(0.0,1.)[b]{$\sim$ } } \end{picture}4}$ ] gev@xmath0 . we point out that out of various sde and bse solutions in the sd - bs approach , once that @xmath144 has been correctly reproduced by @xmath9 [ see eq . ( [ fpi ] ) ] , the transition form factor is most sensitive on @xmath45 , or , more precisely , on its values at small and intermediate momenta @xmath187 , where @xmath45 is still appreciably different from 1 . ( [ ampwtilde])-([approxfactor ] ) already explained how @xmath45 drives @xmath101 upwards through @xmath164 for large @xmath21 in the case of the bc vertex . to illustrate what happens at @xmath16{$<$ } } \put(.7,-0.3){\makebox(0.0,1.)[b]{$\sim$ } } \end{picture}8}$ ] gev@xmath0 when the @xmath45-profile is decreased , let us enforce by hand the extreme , artificial case @xmath188 . ( to avoid confusion , we stress it is for illustrative purposes only , as we can not have such a sd - solution in the adopted approach of refs . @xcite . ) this leads to the curve traced on fig . 3 by small crosses , pertaining to the usage of both bc and mcp vertices ( as well as the cp ones ) , since @xmath188 makes @xmath189 . this curve illustrates how the heights of the curves depicting @xmath101 depend on how much the @xmath45-profile exceeds 1 . obviously , for both the solid curve and the dashed one , the agreement with experiment would be improved by lowering them somewhat ( at least in the momentum region @xmath190{$<$ } } \put(.7,-0.3){\makebox(0.0,1.)[b]{$\sim$ } } \end{picture}4}$ ] gev ) , which could be achieved by modifying the model @xcite and/or its parameters so that such a new solution for @xmath45 is somewhat lowered towards its asymptotic value @xmath191 . ( of course , in order to be significant , this must not be a specialized re - fitting aimed _ only _ at @xmath45 . lowering of @xmath45 must be a result of a broad fit to many meson properties , comparable to the original fit @xcite . this , however , is beyond the scope of this paper . ) by the same token , high precision measurements of @xmath17 ( such as those planned at jefferson lab @xcite ) can be especially helpful in obtaining information on @xmath45 empirically . first note that fig . 3 reveals that in the region around @xmath192 gev@xmath0 we are already in the regime where the dependence on the bs - amplitude @xmath9 is lumped into @xmath144 . therefore , high - precision measurements of @xmath17 in the region close to the asymptotic regime will give information on the integrated strength of @xmath45 although not on @xmath45 itself . however , since it is known that the form of that function must be a rather smooth transition ( _ e.g. _ , see @xcite ) from @xmath193 for @xmath21 near 0 , to @xmath194 in the @xmath21-domain where qcd is perturbative , such measurements @xcite would give a useful hint even about @xmath45 itself namely about what solutions for @xmath45 one may have in sensible descriptions of dynamically dressed quarks and their bound states . the authors acknowledge the support of the croatian ministry of science technology to the conference nuclear and particle physics with cebaf at jefferson lab " , dubrovnik , croatia , november 3 - 10 , 1998 . 10 c. d. roberts , _ nonperturbative qcd with modern tools _ , . roberts and a.g . williams , prog . part . . phys . * 33 * , 477 ( 1994 ) . d. kekez , b. bistrovi ' c , and d. klabuar , int . a * 14 * , 161 ( 1999 ) . m. bando , m. harada , and t. kugo , prog . . phys . * 91 * , 927 ( 1994 ) . c. d. roberts , nucl . a * 605 * , 475 ( 1996 ) . frank , k.l . mitchell , c.d . roberts , and p.c . tandy , phys . lett . b * 359 * , 17 ( 1995 ) . p. jain and h. j. munczek , phys . d * 48 * , 5403 ( 1993 ) . r. alkofer and c. d. roberts , _ phys . * b369 * , 101 ( 1996 ) . d. kekez and d. klabuar , phys . b * 387 * , 14 ( 1996 ) . d. klabuar and d. kekez , phys . d * 58 * , 096003 ( 1998 ) . ( cello collaboration ) , z. phys . c * 49 * , 401 ( 1991 ) . j. gronberg _ ( cleo collaboration ) , phys . d * 57 * , 33 ( 1998 ) . cebaf letter of intent # loi94/005 . spokesmen : a. afanasev , j. gomez , and s. nanda ; a. afanasev , hep - ph/9608305 . m. hayakawa and t. kinoshita , phys . d * 57 * , 465 ( 1998 ) . g. p. lepage and s. j. brodsky , phys . d * 22 * , 2157 ( 1980 ) ; s. j. brodsky and g. p. lepage , _ ibid . _ * 24 * , 1808 ( 1981 ) . r. jakob _ et al . _ , j. phys . g * 22 * , 45 ( 1996 ) ; p. kroll and m. raulfs , phys . b * 387 * , 848 ( 1996 ) ; f .- cao , t. huang , and b. q. ma , phys . rev . d * 53 * , 6582 ( 1996 ) . i. v. musatov and a. v. radyushkin , phys . d * 56 * , 2713 ( 1997 ) . a. v. radyushkin and r. t. ruskov , nucl . b * 481 * , 625 ( 1996 ) ; _ the asymptotics of the transition form factor @xmath143 and qcd sum rules _ , hep - ph/9706518 . v. l. chernyak and a. r. zhitnitsky , phys . rep . * 112 * , 173 ( 1984 ) , and references therein . a. manohar , phys . b244 * , 101 ( 1990 ) . d. kekez and d. klabuar , _ @xmath22 transition and asymptotics of @xmath195 and @xmath196 transitions of other unflavored pseudoscalar mesons _ , hep - ph/9812495 . p. maris and c. d. roberts , phys . c * 56 * , 3369 ( 1997 ) . p. jain and h. j. munczek , phys . d * 44 * , 1873 ( 1991 ) . h. j. munczek and p. jain , phys . d * 46 * , 438 ( 1992 ) . a. bender , c. d. roberts , and l. v. smekal , _ phys . lett . _ * b380 * , 7 ( 1996 ) . v. a. miransky , _ dynamical symmetry breaking in quantum field theories _ ( world scientific publishing co. , singapore 1993 , isbn 9810215584 ) . c. j. burden , c. d. roberts , and m. j. thompson , phys . b * 371 * , 163 ( 1996 ) . j. s. ball and t .- w . chiu , phys . d * 22 * , 2542 ( 1980 ) . d. c. curtis and m. r. pennington , phys . d * 42 * , 4165 ( 1990 ) . p. maris and c. d. roberts , phys . rev . * c58 * , 3659 ( 1998 ) . s. j. brodsky , c .- r . ji , a. pang , and d. g. robertson , phys . d * 57 * , 245 ( 1998 ) . d. v. shirkov and s. v. mikhailov , z. phys . * c63 * , 463 ( 1994 ) . c. d. roberts , _ `` dyson schwinger equations : connecting small and large length - scales '' _ , nucl - th/9901091 , fizika * b * , these selected papers of cebaf - dubrovnik 98 conference . f. j. yndur ' ain , _ quantum chromodynamics _ , springer verlag 1983 , isbn 0 - 387 - 11752 - 0 . n. brown and n. dorey , mod . phys . lett . * a6 * , 317 ( 1991 ) . h. pagels and s. stokar , phys . d20 , 2947 ( 1979 ) . r. jackiw and k. johnson , _ phys . _ * d8 * , 2386 ( 1973 ) . * our results for @xmath101 are presented from @xmath23 up to @xmath129 gev@xmath0 . the data presently exist only for @xmath197 gev@xmath0 . cello data points are circles and those of cleo are triangles . the dash - dotted line represents our @xmath101 evaluated in the chiral and soft limit approximation ( [ softlimbs ] ) and only with the bc @xmath57 vertices . our results _ without _ that approximation are depicted by the solid line for the case of the bc vertex , and by the dashed line for the case of the mcp vertex . this latter one gets lower beyond @xmath198 gev@xmath0 and does not yet saturate , but at higher @xmath21 ultimately must approach asymptotically @xmath199 for higher @xmath21 . this limit is also denoted by the dashed line , but the straight one . the brodsky - lepage interpolation formula @xcite is the dotted curve approaching the ( also dotted ) straight line denoting the pqcd limiting value @xmath107 . the line of empty squares denotes basically the pqcd prediction ( but with fixed @xmath130 ) of ref . the pion distribution amplitudes ( [ genform ] ) are depicted in the dimensionless version @xmath200 for the three values of @xmath126 pertinent to the discussion in this paper . @xmath201 yields the asymptotic distribution amplitude ( the dotted curve ) , @xmath202 pertains to the pion distribution amplitude depicted by the solid curve and corresponds to the large-@xmath21 values of our @xmath17 evaluated with the bc @xmath57 vertex . @xmath203 corresponds to the transition form factors predicted by refs . @xcite and @xcite for the largest @xmath21-values ( @xmath185 gev@xmath0 ) at which @xmath17 has been measured so far . * the comparison of our results for the pion transition form factor ( times @xmath21 ) with the cello ( circles ) and cleo ( triangles ) data . our results for @xmath101 are depicted by the solid line for the case of the bc vertex , and by the dashed line for the case of the mcp vertex . this latter one gets lower beyond @xmath198 gev@xmath0 and does not yet saturate at the presently accessible momenta although approaches asymptotically @xmath199 for much higher @xmath21 . ( this limit is denoted by the dashed straight line . ) the little crosses denote ( for both bc and mcp @xmath57 vertices ) our @xmath101 when we enforce @xmath188 by hand .
we review the pion - photon transition form factor calculated in the schwinger - dyson approach and an impulse approximation . we present results for it far above the scales presently accessible for measurement , up to 36 gev@xmath0 , and demonstrate agreement with the analytically inferred asymptotic behavior , for which we also provide a new derivation . we discuss how measurements at jefferson lab can provide information on how quarks are dynamically dressed . 1.5 cm